The Engineering Dad

3 Years of Foreign Language + Advanced STEM = Table Stakes. What's Your Kid Building?

Tue, 01 Sep 2026 11:32:36 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

If you haven’t read the past College CS Toolkit preview series, check these:
Part 1: Your GPA Gets You into School. Your GitHub Gets You the Internship.
Part 2: Your To-Do List App Is a Rejection. Build These Two CS Projects Instead.
Part 3: What AI Resume Builders Get Wrong About CS Internships
Part 4: How to Master the Interview Process
Part 5: The One Skill That Decides Your Return Offer
Part 6: Your Kid’s Salary Negotiation Fear Is Costing Them Thousands

I am a Tech dad of 3 CS grads (Cornell, and Rice chosen over Caltech/Columbia), so I see this question all the time in parent forums. I also realize the anxiety behind it, too, because I and my wife wondered the same when dealing with college admissions for our children. Because my native language is not English (nor my wife’s), we also wondered if that would make any difference in our children’s college admissions. Ultimately, you are probably stressed about whether 3 years of foreign language is enough for your kid’s CS applications. Based on our real world experience, it is enough.

3 Years Is Enough (But Here’s the Catch)

Top CS and engineering programs do not require 4 years of a foreign language. Most recommend 3 to 4. If your child has completed 3 years of rigorous language study in the same language, that’s a solid credential, and it will meet high school graduation requirements. The admissions office is not going to reject an otherwise strong applicant because they took 3 years instead of 4.

I see this question pop up in parent groups every week. The anxiety comes from the same place where parents read conflicting advice online. One thread says “4 years,” another says “3 is fine.” Nobody in those threads connects the dots to what actually matters for STEM admissions. That’s the gap.

The catch is this. The open slot cannot sit empty. This is where most parents mess up. They see the open slot and think, “Great, my kid has a free period.” Or they fill it with an elective that is not rigorous. Or let the kid take AP Psychology because it’s the only one left that kid hasn’t taken (but this class is usually for freshman or sophomore students) Or they let the kid coast. That’s how the application gets hurt.

The Replacement Matters More Than the Language Requirement

If you drop the 4th year of language, then that slot needs to be Multivariable Calc, Linear Algebra, Differential Equations, or post-AP science, something that is above beyond the typical high school classes. These students may have already completed AP Calc BC, AP Physics C already. These are the baseline courses now for competitive STEM admissions.

My own children all did this. They completed their language requirements by junior year (three years, same language). Then they moved into advanced mathematics and physics. Multivariable Calc, Linear Algebra, Differential Equations in some cases. All three of them got accepted into top CS programs. This strategy works.

Why? Because top CS programs care about mathematical depth. They want to know that your kid can handle proof-based mathematics, not just computational fluency. Linear Algebra is where engineering meets abstraction. Multivariable Calc opens the door to physics and systems thinking. These courses signal readiness for the actual work of CS.

Taking a 4th year of basic foreign language signals nothing about your child’s readiness for CS. Taking Multivariable Calc signals everything.

As I covered in Why Dropping French 3 Isn’t a Dealbreaker for Engineering Applicants, the key insight is this. “That rigid advice is missing a massive piece of strategic context.” Parents hear “4 years is best” without understanding the trade-off rule. Surely, those parents whose kid got accepted to a top school and has taken 4 yrs of foreign language will tell you, “You need 4 yrs of FL.” Whether this is a must or not tells a different story.

♻️ Restack if you know your friend who may benefit with this info.

Table Stakes vs. Real Differentiation

That combination, 3 years of language plus advanced STEM coursework, is now table stakes. Most competitive applicants have it. So the kid who took 3 years of language and crushed Multivariable Calc is meeting the floor, not standing out, either.

This is hard to hear, but it’s true. The coursework alone will not move the needle anymore. Top-tier STEM applicants now all have the same transcript. AP Calc, AP Physics, honors languages, advanced electives. Everyone. So what makes the difference now?

Here’s the paradox. Once you’re in the “top tier” of competitive applicants, and this coursework gets you there, the ranking of your school matters far less than you think. As I noted in The CS Showdown, “Once you are in the top 50 of ABET-accredited engineering programs, the ranking number is noise.” The same principle applies to credentials. Once everyone has the credentials, the credential stops differentiating.

From Consumer to Builder

I see parents spending the whole summer writing college essays, stressing over test scores, optimizing their course list, trying to find the perfect online class to list on the application. Meanwhile, their kid is collecting credentials instead of building artifacts. They are running the consumer playbook, what is check the boxes, stack the credentials, hope it’s enough.

As I outlined in Pre-Summer Action Plan, “With school ending in weeks and summer approaching, now is the moment to transition from passive consumption to active building.” The window to make this shift is now, before senior year starts.

Top CS programs want to see what your kid makes when the classroom ends. Not another certificate or credential. An actual deployed thing that they can see. A web app with real users that they can verify. A competitive robotics project that they can watch. Open source contributions that they can validate. A tool or system they built and shipped that they can play with. Hardware projects to see all details. Research contributions that others depend on. Anything that shows proof of work, not just proof of credentials.

I know a student who got rejected from three Ivy League CS programs with a 1550 SAT, 4.0 GPA, and perfect AP scores. Why? Because the application showed a consumer, someone who optimized every metric and had nothing to show for it except credentials. He had no projects, no code and no artifact admission officers could actually see.

This is exactly what I described in Confessions of a Tech Hiring Manager: “Your Child Has a 4.0 GPA. So Does Everyone Else. Here is What He Needs Instead.” The GPA becomes invisible once everyone has it.

I also know a student from a public high school who got into Cornell CS and Rice CS with slightly lower test scores because they had built and shipped a tool that was actually used by real people. Hundreds of GitHub stars existed. A small community around it that was actively trying to improve the code. That’s what moved the needle.

As I explored in The Builder Delta, “The gap between those two sentences is where most applications die.” The gap between listing a club and building something real. That gap is where admissions officers look hardest.

The language requirement is settled now. The math coursework is table stakes. The differentiator is what your kid builds outside the structured curriculum. That’s where the real evaluation happens.

What You Can Do Today

Do this. Confirm the language decision (3 years is enough). Don’t waffle on it anymore. Then sit down with your kid and ask, what are you going to build?

Not a class. Not a certificate. An actual thing, a deployed web app, a robotics project, open source work, or a shipped tool that the child can proudly show and tell.

If your kid is interested in hardware or systems, competitive robotics (FTC, FRC) is the clearest path to a builder portfolio. If your kid is interested in software, a deployed tool or app is the signal. If your kid is interested in research or AI, consistent open source contributions show proof of work.

The language decision? That’s settled. You can move on.

The real question is the one I’d be asking myself right now. Does my kid have 12 months to start building something real, or are we still stuck in the consumer game?

That is how you actually stand out.

Let’s build.
Masaru

P.S. If you are still asking “should my kid have a portfolio site?” the answer is yes, probably. Not required, but the advantage is huge. Even a simple one. Even if it takes six hours. Wix, Squarespace, Google Sites, any platform works if it’s clickable and shows your work.

If your child is a high schooler, The High School CS Internship Playbook I put together on Gumroad walks through the high school level “how to get internship” process step by step, including the Builder Project framework, the portfolio review process, and real examples from students who were at this exact crossroads. You can find it here.

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

Merit Scholarships Do Exist. You are Just Looking in the Wrong Schools

Thu, 27 Aug 2026 11:43:29 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

Pick Your Merit Battle

Your son has a 1540 SAT, a 3.93 unweighted GPA, weighted even higher. Top percentile nationally. The college counselor said aim high. Your Facebook group is full of parents with similar stats posting about their kids getting into top 25 schools. So you expected merit scholarships to follow. Instead, the financial aid letters came back the same way for all of you. Full sticker price. $85,000 a year. Maybe a small need-based grant if your family qualifies.

Your son will get a merit scholarship. Merit scholarships simply don’t exist at Stanford, MIT, or the Ivy Plus tier where his 1540 sits exactly where everyone else’s sits. But at schools where 1540 makes him exceptional instead of typical, he’ll have real leverage. That’s where merit funding actually concentrates.

Why Elite Colleges Don’t Offer Merit Scholarships

I’ve been in tech hiring for 20 years. I know how supply and demand works. When you have a luxury car dealership with 10 inventory slots and 150 people willing to pay full price for one of those cars, you don’t negotiate the sticker price down. You won’t discount for someone with a strong credit score.

That’s exactly what’s happening at T25 schools. The acceptance rate is 4 to 8 percent. Tens of thousands of applicants have 1540+ SATs and near-perfect GPAs. Many of them will pay full price out of pocket or take loans. Schools have zero financial incentive to give away institutional merit scholarships to students who will attend anyway because they got into their dream school. That’s how elite college funding actually works.

♻️ Restack if you know your friend who may benefit with this info.

Thanks for reading! This post is public so feel free to share it.

Where Merit Actually Concentrates

Most elite private institutions in the Ivy Plus tier don’t award institutional merit scholarships at all. They offer need-based financial aid only. If your family makes too much to qualify for need-based aid, you’re paying $85,000 a year.

Merit scholarships concentrate at schools with different economics. These are schools where your child’s profile sits comfortably in the top 5 percent of the applicant pool. I mean the top 5 percent of applicants to that specific school, not nationally. Your son’s 1540 sits at the median at Stanford. At a different school, it might sit in the top 5 percent. That’s where the leverage appears.

My second child received $25,000 per year in merit scholarship. My third child received a trustee scholarship at Rice that made the education a full ride, combined with need-based financial aid. Both outcomes came from strategic targeting. We identified schools where their stats created actual leverage, not just the lottery ticket you get at T25 schools. As I cover in Winning the “Affordability Gap”, the May 1st deadline isn’t just a decision day, it’s the final minute of a high-stakes negotiation. You need leverage to negotiate, and leverage only exists where your child’s profile is rare.

Run the Net Price Calculator on fifteen schools. Most will show the same brutal number, but three or four will be different. Those are the schools where your child’s numbers are rare in the applicant pool, where you actually have leverage.

The Three Realities Parents Need to Accept

First reality. Few schools offer automatic merit scholarships based on test scores. Many state flagships will give an automatic $5,000 to $15,000 for a 1500+ SAT. That amount doesn’t cover $40,000+ annual tuition. When you see a parent bragging about their kid’s 1540 and perfect GPA, assume nothing. Merit aid may not materialize, or it caps at one semester.

Second reality. External scholarships have acceptance rates lower than Harvard. The Coca-Cola scholarship, the Cameron Impact scholarship, the Elks scholarship, these exists. But they receive thousands of applications from students with identical credentials to your child. The acceptance rate hovers around 1 to 2 percent. Even if your child wins one, most cap out at $2,000 to $5,000. That’s helpful money for one year, not four.

Third reality. High test scores are table stakes at T25 schools, not differentiation. A 1540 SAT at Stanford is the baseline median, not exceptional. Your child needs more. They need a demonstrated research project, or a published paper, or a portfolio that proves they built something real. And even with all of that, they’re still in a lottery pool. The merit funding doesn’t follow because the school doesn’t need to offer it. As I outlined in The Builder Delta, the gap between listing “Club President” and describing what that role actually accomplished is where most applications die. High stats alone leave the same gap. I know this sounds conservative to families hunting for the Ivy Plus tier. But here’s what my own kids’ applications taught me.

The Strategic Miscalculation

Families chase prestige out of state to a top 25 school. Tuition is $85,000+ per year. They’re counting on merit scholarships to cover $50,000 or $60,000 of the bill. They’ll apply to a few external scholarships as backup.

That strategy assumes merit funding is a consolation prize at selective schools. It’s not how it works. Merit funding is the primary mechanism at different schools entirely. Schools where your child’s profile is strong enough that the school is competing to attract them, not the other way around.

If merit funding is the goal, the leverage always lies in targeting high-caliber schools where your child sits comfortably in the top 5 percent of the applicant pool. As I covered in The CS Showdown, once you’re in the top 50 ABET-accredited engineering programs, the ranking number is noise. What matters is fit, pipeline, and where your child’s profile creates competitive advantage. High-caliber schools have excellent engineering programs, strong exit outcomes, and great communities. They’re just not the schools with 4 percent acceptance rates. When your child is in the top 5 percent, they compete for the school’s attention, not the other way around. That’s a position of strength, not a compromise.

What to Actually Do

Run the Net Price Calculator on each school you’re considering, looking at actual out-of-pocket cost after aid, not sticker price. Do this on fifteen schools. Watch the numbers. Most will sit between $80,000 and $85,000. Three or four will drop to $40,000 or $50,000 per year, and that’s where merit is concentrating. As I explain in The “Trojan Horse” in Your Financial Aid Letter, understanding how to read the fine print of an offer letter is critical, schools hide real costs in the details.

Identify schools where your child’s profile sits in the top 5 percent of admitted students. Your school counselor can help calculate percentile within each school’s applicant pool. Look for schools where your child is in the top 5 percent AND where the program actually fits their goals and interests. When your child’s scores sit there, merit scholarships follow.

Parents in Facebook groups want to hear that their child got into an Ivy. I get it. I and you both want our kids to have options. But an $85,000 annual bill for four years is $340,000 in obligation. That constraint shapes every decision. The standard advice of chase prestige, pad the resume, aim for the name brand is broken. As I wrote in Why “Summer Camps” Don’t Work, that advice costs more and prepares your child for admissions theater, not for a career.

My kids got merit aid at Rice, not Stanford or MIT or Princeton. They got merit scholarships at schools where their profiles were rare and schools competed for their enrollment. That’s where the real leverage lives. Find those schools by running NPC on your list, calculating where you sit in the top percentile, then negotiating from data and percentile ranking, not hope. That’s how my kid found merit aid at Rice. That’s how you will too.

Let’s build such a profile for your child.
Masaru

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

Your Kid's Math Score Is Not What Matters Anymore

The Engineering Dad — Tue, 25 Aug 2026 11:40:39 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

Math Curriculum as Consumer Thinking

A parent on Facebook asked whether her 6th grader should take Algebra 1 now or stick with Pre-Algebra to stay with his grade. He is testing strong in math, he is self-studying Art of Problem Solving, and he wants top 20 CS. She is worried about logistics. Will a 0-period class mess up his schedule? Will an online course create too much load?

Here is what I would ask instead. For the last decade, the answer was straightforward. Math rigor was a differentiator. You accelerated hard through high school, you cleared the calculus sequence by junior year, you went to MIT. Math was the thing that set you apart.

That is not true anymore.

Math rigor is now table stakes.

Not a differentiator.

♻️ Restack if you know your friend who may benefit with this info.

What Table Stakes Looks Like

My three kids all went CS to top 20 schools. Cornell, Rice, and now one pursuing an AI/ML master’s at UIUC. And their math paths were wildly different.

My first born took college level Calc I, Calc II, and Differential Equations in her junior and senior year at high school. She went to Gatton Academy in KY after her sending school did not offer AP in freshman and sophomore. She took college courses through WKU.

My second child took AP Calc BC and AP Physics C in sophomore year. Then Calc II, Calc III, and Linear Algebra by graduation. He came in hot and stayed aggressive.

My third child started with AP Calc BC and AP Physics in freshman year. Then AP Chem, AP Stats in sophomore. Calc 3 junior, Differential Equations senior. She is the outlier in acceleration intensity. This might be an expected sequence of courses because we learned through our 1st and 2nd children, and the sending high school also started to offer advanced math classes as the time went on.

Three different timelines, three different paces, one outcome that mattered. All three had rigorous math by 11th grade across different entry points, and all understood calculus sequences before senior year. Math was settled by the time college started.

But their friends accelerated just as hard. Same AP Calc BC in sophomore or junior year. They ended up at University of Kentucky, Purdue, solid state schools. But not Rice. Not Cornell. Not Northeastern (NEU)

The math did not separate them. Something else did.

The Invisible Student

I coached a student years ago who tested sharp in 6th and 7th grade math. Legitimately strong. Then he became invisible.

Not held back by teachers. He just progressed in a different pace after 7th grade.

He was capable, but talent lived elsewhere. And because everyone around him was accelerating, he felt the pressure to keep accelerating in the wrong direction. Eventually, he pivoted to pre-med instead of CS. He is in medical school now, thriving. But here is the point. Early math promise does not guarantee anything. That invisibility is the trap. Talent that does not translate to output does not matter.

As I have covered in my analysis of medical school admissions, a generic credential is no longer a magical shield. The same applies to math in CS. You can be strong at math and still have nothing to show for it.

What Actually Moved the Needle

I spent years at Cincinnati Children’s, Toyota, Intel and HP in systems, product and technical program management. I have reviewed resumes from kids with perfect math backgrounds. Perfect test scores. Crystal clear transcripts. The resume I stopped at had a YouTube link showing a real project. A product demo.

Resumes with titles got rejected fastest. Club President. Team Lead. Chapter Founder. No context. No proof that anything changed because they held that role.

The gap between those two sentences is where most applications die.
— The Builder Delta: Why “Club President” Isn’t an Artifact

After their math got locked in, my three kids had time to build something verifiable. My first born ran a research project at Gatton Academy and Cincinnati Children’s Hospital during high school, submitted work for publication, and won the state science fair. My second child built FTC robotics systems and created his own Minecraft server managing his friend group online, competing at the FIRST world competition multiple times. My third child built FTC robotics, competed at FIRST worlds multiple times, created a portfolio site, and started a paid internship from 9th grade working on client coding projects.

That is what separated them from their peers. The applicants who got in had finished calculus by 11th grade. But here is what actually decided admission, the research project, the GitHub profile, the hackathon entry, the robotics competition, the portfolio that worked. The verifiable build.

How to Actually Think About This

If your 6th grader is strong in math, accelerate. Accelerate because it frees up time.

Get the math sequence locked down by 11th grade. Algebra 1 in 6th grade, Calc BC in 10th, community college courses, or a mix of online and in-person. The path does not matter. The timing does. By junior year, math should be off your plate.

With school ending in weeks and summer approaching, now is the moment to transition from passive consumption to active building.
— Pre-Summer Action Plan: From Consumer Trap to Builder in 8 Weeks

Then what does your kid build in that freed up time?

I mean code. A project that lives somewhere public. A hackathon entry. A research paper submitted. A GitHub profile showing iteration and complexity. A portfolio site. A robotics team. Something a stranger can evaluate without asking for context.

That is the differentiator now.

The 6th grader accelerating to Algebra 1 while his friends stay in Pre-Algebra? That advantage lasts maybe two years before everyone catches up or gets there by a different route. But the 10th grader who has a hackathon win or an open-source contribution or a published research project or a portfolio site with paid client work while his peers are still doing homework? That is the kid who stands out.

Once a student builds their first hardware-software bridge through robotics, research, or client work, they have paid their technical debt. Every project after that becomes faster, more complex, and higher-signal to admissions officers. Your kid locked in math by 11th grade, checked. Your kid with a verifiable build, a competition entry, a GitHub repository, published research, a portfolio that works, checked. That is the admission package.

Let’s build.
Masaru

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

Don't Choose Between Biology and Engineering. Here's the Third Door Nobody Tells You About

The Engineering Dad — Thu, 20 Aug 2026 11:22:42 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

Engineering Major is Not Really Equal to Biology Major

A couple weeks ago, a parent asked me something I hear more often than you’d think. “My kid loves both biology and engineering equally. The Common App is asking for intended major. Which should they apply as?”

The answer most people give is wrong. They’ll say “follow your passion” or “it doesn’t matter, you can switch later.” But it does matter. The real pathway isn’t a choice between them.

Here’s what I’ve seen work repeatedly. When students start in engineering with a genuine biology interest, they have access to something most kids never get.

A third door.

Not pure biology. Not pure engineering. But biomedical engineering. Bioinformatics. Biological engineering. The interdisciplinary middle where the quantitative rigor meets life sciences.

The kid who loves both doesn’t need to choose. They need to start with the hard major.

Why Engineering First Opens the Path

The transfer direction matters more than the starting major. I’ve seen this play out dozens of times. A student starts in engineering with biology interests. Takes the freshman math sequences, organic chemistry, engineering coursework. By sophomore year, they are ready to pivot into a bioinformatics concentration, or a biomedical track, or biological engineering. The engineering foundation is already there. The doors are open.

Now reverse it.

A kid starts in pure biology hoping to add engineering later. They miss the linear algebra. They miss the computational systems thinking. The engineering school looks at their transcript and says “you don’t have the prerequisites.” That’s not because they are not smart enough. It’s because the structural barrier exists. Parents must remeber getting into engineering after starting in biology is harder than getting out of engineering into bio.

As I covered in The CS Showdown: How to Choose Between Top Engineering Schools (Without Looking at Rankings), once you are inside a top engineering program, the ranking matters far less than the pipeline and personality of that specific school. The same principle applies here. The entry door matters more than the ranking of the school offering it.

So the paradox is this. If your kid is torn between both, start with the one that’s harder to enter. That’s engineering.

♻️ Restack if you know your friend who may benefit with this info.

What If They Hate Engineering?

Now, you may be thinking… What if my kid gets into engineering and hates it after a semester? That’s a fair question I should answer. Here’s what I’ve seen. When a student struggles early, it’s usually not the major itself. It’s that they haven’t found the specific application area yet. I hired a woman once who majored in astronomy. Wanted to work in space exploration. But her first job out of college was database administration. Different field entirely. She struggled at first, catching up with all the engineering disciplines she’d skipped. But once she focused, really focused on one problem space, she adapted. She became excellent. The major didn’t lock her in. The focus did. Your kid can do the same thing.

Start in engineering.

Give it a semester or two. Then focus on the specific application that actually calls to them.

As I covered in Confessions of a Tech Hiring Manager: Why I Ignore 4.0 GPAs, what actually separates strong candidates from the rest is demonstrated ability and the willingness to focus on depth. That same principle holds here. A student who starts in engineering and then focuses on bioinformatics or biomedical engineering shows exactly that kind of thinking.

What This Actually Looks Like

When a student does this right, they’re not sacrificing biology. They’re building a foundation that lets them be more specific about which biology work they actually want to do.

I coached a student a few years back who was genuinely split between them. She loved the precision of engineering work, the problem-solving rigor. She also loved thinking about applications in healthcare. By junior year, she was in a biomedical engineering track, doing capstone work on medical device design. She wasn’t less of a biologist because she started in engineering. She was more of a systems thinker because of it.

“Students do not need 2,000 problems. They need to understand these 75 and develop the ability to apply the concepts.”
— The AI “T-Shaped” Major: Why Ross and Kelley are Winning the B-School Arms Race

That’s the payoff. Your kid gets the rigor of engineering coursework plus the application area they actually care about. They become a T-shaped student. Deep in one area. Broad across systems thinking. That’s what companies actually hire.

The Common App Mechanic Nobody Explains

Most parents get stuck right here. When my 1st born was applying, we filled out the Common App and I thought we’d picked the major. That was it. Final answer.

That’s not how it works.

What actually matters, what admissions officers actually read, is what you list on each school’s supplement, and you can tailor that by school.

If your kid is genuinely strong in math and physics, list engineering at the schools where that profile stands out most. If the school’s engineering program is your reach and your kid’s test scores and competition experience are very strong in math, that’s where you want to emphasize engineering.

Then at the schools where your kid’s biology research or science project work is the strongest piece of the application, list the biology-adjacent major on that school’s supplement. Biomedical engineering. Bioinformatics. Biological engineering. Whatever that school offers.

This isn’t gaming the system. It’s alignment. As I covered in From Science Fair to Salary: How to Turn High School Research into a Paid College Internship, how you position and package your work is just as important as the work itself. The same applies to your intended major. You’re not lying about your kid’s interests. You’re showing each school the part of your kid’s profile that best matches what that school actually cares about.

The Real Strategy

So here’s what I recommend. If your kid loves both biology and engineering, apply to selective engineering schools where your kid’s engineering background is strongest. Apply to schools with strong biomedical or bioinformatics programs where your kid’s biology project work shines. Apply to a mix.

On the main Common App, you don’t have to box yourself in. The career goal is aspirational and non-binding.

On each school’s supplement, tailor the intended major to the part of your kid’s profile that’s strongest at that school. It’s strategic. It’s honest. And it keeps both doors open.

As I covered in Why “Looking for an Internship” is the Worst Way to Find One (And What to Do Instead), what separates students who land opportunities from those who don’t is the proactive, intentional approach. That same mindset applies to your major strategy.

Don’t passively list the first major that comes to mind.

Be intentional.

Be strategic.

Show each school why your kid is the right fit for that school’s specific program.

Then, once your kid gets in, and they will get in somewhere, they start in engineering if they want the optionality. Because going from engineering to biomedical engineering or bioinformatics is straightforward.

The reverse is not.

Your kid doesn’t have to choose between biology and engineering right now.

They just have to choose the entry door that keeps both options available on the other side. And if you’re trying to figure out which schools even have those bioinformatics and biomedical engineering programs, or which schools’ engineering programs match your kid’s specific strength? Leave a comment. I can help you.

Let’s build that school list.
Masaru

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

The Spanish Major Trap: Why Narrative Mismatch Kills Admissions

The Engineering Dad — Tue, 18 Aug 2026 11:27:19 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

Red flag = Narrative Mismatch

A student did hospital volunteering, published in medical journals, worked clinic hours. Then applied as a Spanish major. Admissions officers immediately asked, why?

The red flag is the narrative mismatch, not the major itself. When your activities don’t line up with your stated major, AOs sense strategy over authenticity. Unless this student has a deep, multi-year immersion in Spanish language work, community advocacy, or linguistics research, the combination reads manufactured. And T20 AOs spot this strategy instantly.

The student built their entire record around medicine including hospital hours, medical publishing, clinic work. All of it points to one direction. Then they try to apply as a Spanish major. The AO flips through the application with one question, as you may also as the same, if you care this much about medicine, why Spanish? If you care this much about Spanish, why all the medical activities? The answer is usually strategy. And strategy reads as inauthentic.

As I covered in The Builder Delta: Why “Club President” Isn’t an Artifact, the gap between what you list and what you actually built is where most applications die. Credentials without a coherent narrative don’t move the needle.

♻️ Restack if you know your friend who may benefit with this info.

Premed students lock themselves into one track. Your activities must be medicine. Your courses must be biology, chemistry, or biochemistry. Your summer must be clinical or medical research related. Build a record that doesn’t align with those three fields, and you have officially created a gap that needs explaining. That’s where the Spanish major trap happens. You are trying to solve a problem that engineering students don’t usually face.

Here is why. My 1st daughter started at Cornell as a biomedical engineering major. She was exploring the MD route and was a BS/MD candidate applying to multiple BS/MD programs. She volunteered at hospitals, did research intern as a high schooler at the nation’s one of the leading Children’s hospitals, took the premed prerequisites, built what looked like a medical trajectory. Sophomore year at Cornell, she realized medical field dream was no longer pulling her in. She was staying up late building Python code instead. So she switched to computer science.

Cornell lets you choose your major sophomore year. That is the difference. Engineering is a foundation, not a locked path.

As I discussed in The CS Showdown: How to Choose Between Top Engineering Schools, once you are in the top 50 of ABET-accredited engineering programs, the specific ranking is noise. What matters is the flexibility built into the curriculum. Biomedical engineering teaches physics, circuits, materials, computation, and biology. Each one opens doors to medical school, to computer science, to robotics, to finance. The major doesn’t close paths.

It opens them to multiple directions.

This is huge.

This is exactly why I have been supporting engineering majors so that you will be able to get a job after college graduation with the right approached preparation.

Here is the key that many parents miss. You can take pre-med coursework inside an engineering major. You don’t have to choose between exploring medicine and keeping your options open. The engineering major provides cover for that exploration.

Now contrast that with what would have happened if my child had started as a biology major. She would have had to retake courses, rebuild her transcript, and explain why she had abandoned the biology track. The major and her story would have been locked together from day one.

Lost credits.

Lost time.

Harder to explain to colleges if she had wanted to transfer.

Engineering gives you a two-year window. Premed gives you one semester before organic chemistry locks you in. After that first semester, your prerequisites chain forward. Backing out means lost credits and a disrupted timeline.

That is the asymmetry of concerns. Engineering students can explore medicine, finance, software, robotics, then commit to one. Premed students commit first.

I advised another student who started biochemistry thinking she wanted medicine. Year two, she realized she wanted software engineering instead. The biochemistry credits didn’t count toward the CS major.

She lost a year rebuilding. This is very common.

Why? Because if you are a premed major, your major and your story are the same thing.

Chemistry major means “I want medicine.”

Biology major means “I want medicine.”

Biochemistry major means “I want medicine.”

Build a record outside that story and AOs see a mismatch.

Engineering is different.

You are an engineer.

You can do medical work, software work, robotics work, business work, all of it fits under one major that makes sense.

As I outlined in Why Dropping French 3 Isn’t a Dealbreaker for Engineering Applicants, rigid advice about credentials often misses a massive piece of strategic context. The same principle applies to majors. Insisting that your activities must match a specific major is rigid thinking.

Engineering majors don’t operate that way.

If your rising junior loves biology and is considering medicine, start by asking this today.

Does your child want to explore medicine, or does your child want to commit to medicine?

Those are different decisions.

If they are 16 or 17 and still figuring it out, engineering gives them space to explore without committing. If they are certain medicine is the path, a pre-med major is fine. But certainty at 16 is rare. Most students change direction at least once before graduation. They are rarely certain at that age.

The real edge in engineering admissions is not the major itself. It’s the flexibility built in. You can change your mind without changing your story. You can explore multiple paths without creating a narrative that AOs have to untangle.

If you are tempted to use a major as strategy, to look different in an overcrowded applicant pile, remember this. As I covered in Pre-Summer Action Plan: From Consumer Trap to Builder in 8 Weeks, a logical question built on a flawed premise never works in admissions. Using a major as a cover story is exactly that, strategy standing in for authenticity.

AOs have read thousands of applications. They know when your major doesn’t match your record.

In my opinion, the mismatch is not clever differentiation.

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

Your Kid's Resume Is Not the Problem. The Portfolio She Does Not Have Is.

The Engineering Dad — Thu, 13 Aug 2026 11:45:39 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

What is a good resume for high schoolers?

I saw a thread in a parent group last week where someone asked “Are these good resume examples?” and someone else replied “You let ChatGPT create one for you.” That comment made me realize what’s missing from how parents think about resumes.

Here’s what I tell parents. Stop polishing the resume. The real competitive move happened when another student built a portfolio site and started linking to it. Your kid isn’t competing against a perfectly formatted PDF. They are competing against that.

I’ve been hiring for 20 or so years, Intel, HP, Toyota, the usual path, and I can tell you what I stopped looking at first. Resumes with club president titles and no outcomes. Resumes with “hardworking individual” as a bullet point. Resumes that passed the ATS scanner (the machine that filters applications) but made me ask “what actually did this student do?”

The students who stood out were not the ones with pristine formatting. They were the ones who gave me something to click on, a GitHub link with actual code, a YouTube video showing a project running, or a portfolio website that presented their work with images. And if I’m being honest, the portfolio site is the move fewer students think to make. But it’s the one that changes how a hiring manager or admissions officer reads your resume.

My third child did exactly that when she applied to college. She took the resume structure (ATS-safe, keyword-dense) and linked it to a portfolio site that had images and a narrative of her actual projects. Readers could go there and see the full story (what she built, why it mattered, what she learned). The resume itself was clean and scannable, but it was the portfolio site that actually moved the needle. This is The One Skill That Decides Your Return Offer applied to the resume level, make it effortless for the human to understand what you did.

The Two-Gate Problem

A resume has to pass two completely different gates, and most students only think about one.

Gate 1 is the machine scanner, the ATS tool that filters applications for keywords. It’s looking for certain words, certain formats, certain structures that signal relevance. If you are not keyword-dense enough, the machine filters you out before a human ever reads your resume. That gate exists as the first line of defense on their hiring side, and you have to pass it.

Gate 2 is the human. Once your resume lands in front of a human reader (an admissions officer, a hiring manager, a program director), they are not reading for keywords. They are reading for specificity. They are asking “what actually happened here?” And if your resume says “demonstrated teamwork” or “hardworking individual” or “engaged in competitive programming challenges,” the human reader has nothing to work with. Generic claims don’t move people. Specificity does. As I covered in Stop Learning to Code and Start Building, they want proof of real work, scoped problems, solved constraints, actual shipping. Not summaries of learning.

Most high school resumes win Gate 1 and lose Gate 2. They get past the machine. Then a human reads it and has no idea what they actually accomplished.

What Gate 1 Looks Like (The Resume Examples)

Let me show you what I mean. I looked at three recent resumes from parents asking for feedback. All three are from real applications. Here’s the breakdown.

West Coast Student: Passable, But Incomplete

This resume does quantify things like”500+ hours of rehearsals,” “$100,000 of equipment.” That’s the right move. The format is clean, and there’s no wasted space on an objective statement (good call).

But here’s the flaw I noticed, at the bottom, there’s an “Academic Goals” section. “Aspiring to pursue a law degree. Interested in aviation law.” A high school resume shouldn’t be speculating about the future. It should prove what you’ve already executed. That section adds nothing and uses up space that could show actual work.

Verdict is this. passes ATS, mostly fine for Gate 1, but doesn’t move a human reader past “okay, this student did things.”

Classical Academy Student: AI-Fluff Trap

This one has all the classic tells. An objective statement at the top. “Responsible and hardworking student, eager to learn.” That’s space wasted on speculation, and it’s the kind of thing ChatGPT generates when parents try to help.

Then the bullets. “Demonstrates teamwork, discipline, leadership.” These are claims, not proof. Who doesn’t demonstrate teamwork? Who says they lack discipline? These words pass an ATS keyword search for “leadership” or “teamwork,” but they don’t tell a human reader anything.

There’s also redundancy. CPR certification listed twice (once under Certifications, once under Skills), and coursework padding. Weight Training and Drawing 1 sit next to “AP Biology” like they belong in the same category. If you as a human reader, take a quick look, you realize they don’t belong in the same category and something appears “off”.

Verdict is this. It might pass Gate 1 on keyword density, but it reads as AI-written or parent-written, and that kills credibility immediately.

Midwest High School Student: Paper Tiger STEM

This one actually has strong academics. Linear Algebra. Calculus III. USACO (that’s USA Computing Olympiad, serious work). The student has elite coursework.

But then the bullets don’t match the level. “Engaged in competitive programming challenges.” That tells me nothing. What division did they reach in USACO? Bronze? Silver? Gold? Those matter. There are different levels of difficulty. If you reached Silver, say so. If you reached Gold, that’s the bullet.

Then there’s “Public Relations Member.” This is exactly what I mean when I talk about The Builder Delta: Why “Club President” Isn’t an Artifact, the gap between listing a title and showing what actually changed because you held it.

And in the skills section. “Quick learner.” “Strong persistence.” These are soft skills masquerading as technical skills. Don’t list them alongside Java and Multivariable Calculus. They dilute the real technical credibility.

Verdict is this. Strong academics, but the resume is passive. It fails Gate 2 completely because it has no evidence of applied work.

♻️ Restack if you know somebody who may benefit with this info.

What Gate 2 Looks Like (The Portfolio Move)

Portfolios don’t have to be complicated.

My third child’s was a simple Wix site. Four pages. Images. Short descriptions. Took her about six hours to build. But that six hours moved her from “student with credentials” to “student with proof.”

When admissions officers clicked that link, they could see a robotics project with pictures of the robot and a video of it competing. A coding project with a GitHub link and a deployed app. An internship project with artifacts and outcomes. A leadership role with actual metrics. Not summaries. Not claims. Proof.

And here’s the bonus. That same portfolio site worked for internship applications, job applications, anything that needed to answer “what actually did you build?” Most people worry about time. Six hours isn’t time. That’s an investment that pays for four years of applications.

The Action + Context + Metric Rule

If you want both gates to open, forget soft skills. Follow this instead.

Eliminate objectives and soft skill lists. Cut “hardworking.” Cut “quick learner.” Cut “demonstrates teamwork.” Those words pass an ATS scan and lose a human reader. Show, don’t tell.

Quantify results. Don’t say “engaged in competitive programming challenges.” Say “achieved USACO Silver Division competing in C++.” Don’t say “public relations member.” Say “managed social media for the robotics team, growing followers from 200 to 1,200 over one season.” Specific. Metric. Proof. As I covered in Your GPA Gets You Into School. Your GitHub Gets You the Internship, the principle is the same: the metric proves the claim.

Show applied execution. Highlight tangible projects, code repositories, leadership with measurable outcomes, or physical builds. Not coursework. Not titles. What did you ship? What changed because you did it?

That’s it. Action (built, achieved, managed, designed). Context (what it was for, what level, what platform). Metric (the number, the outcome, the specificity that proves it).

What This Means for Your Kid’s Resume Right Now

If your student’s resume has an objective statement, delete it. If it has soft skill claims without metrics, replace them with specific outcomes. If the bullets are vague (”engaged in” or “participated in”), rewrite them using the Action + Context + Metric rule. If there are no links to proof (GitHub, portfolio, demo), add one.

Then run a test. Send the resume to one person who doesn’t know your kid, a mentor, a coach, someone who hires. Ask them. “What did this student actually do?” If they can’t answer in specifics, the resume failed Gate 2. That’s the signal to add the portfolio link and rebuild the bullets.

The resume gets you past the machine. The portfolio gets you past the human. Most students optimize for one and lose the other.

The ChatGPT Problem

One more thing. I saw that comment about ChatGPT in the parent group. Here’s why it matters.

If a parent creates a resume for their student, even a good one, and that student can’t defend it or explain it, admissions officers and hiring managers will know. The credibility breaks immediately.

If a parent uses ChatGPT to structure a resume and then their student can’t elaborate on the bullets during an interview, the story falls apart.

The resume is you talking about yourself. It has to sound like you. It has to reflect what you actually did. If it doesn’t, humans sense it. Machines might not. Humans will.

If You Want a Resume Review

If this is your kid’s situation, resume passes ATS, fails the human, I can help. Reply to this post and we can look at the specific bullets, identify where the portfolio link should go, and rebuild it using the Action + Context + Metric framework.

Let’s build.
Masaru

P.S. If you’re still asking “should my kid have a portfolio site?” the answer is yes, probably. Not required, but the advantage is huge. Even a simple one. Even if it takes six hours. Wix, Squarespace, Google Sites, any platform works if it’s clickable and shows your work.

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

The Complete College CS Toolkit: How to Go From Freshman to Internship Offer in One Summer

The Engineering Dad — Tue, 11 Aug 2026 11:22:19 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

What’s at stake?

College freshmen. You are here. Classes start very soon. Are you ready?

The next three months will be critical for your college career.

It’s honestly too late if you wait until junior year to think of summer internship. The competition starts now. You are competing against students who started building in their freshman summer. The headstart is all you need, but you must start now. Internships may lead to return offers if you frame it right. You are not only just maximizing a summer, but also you are compressing years of career progression into one.

I built the College CS Toolkit because I watched this pattern with my own three kids. They are all CS graduates. Cornell Engineering, Rice (chosen over Caltech and Columbia). Northeastern (NEU) to State Univ (undergrad) to UIUC Master’s in CS with AI/ML concentration. They took less emphasis on grades, but they optimized for what the most important action is, which is “internship sequence”.

Part 1: Your GPA Gets You Into School. Your GitHub Gets You the Internship.

GPA is necessary. It opens doors. No GPA, no interview.

But here’s the trap. Most students think GPA solves the problem.

It does not solve at all these days. GPA is firmware. It’s the baseline. Sufficient for admission. Insufficient for hiring.

Hiring managers don’t care about your GPA. They care about shipped work. Real, deployed, public projects with URLs they can click.

Your GitHub is your internship portfolio.

This is the identity shift that changes everything. Stop thinking like a student, start thinking like an engineer. Students optimize for grades. Engineers optimize for building things that work.

The Logic Gate is this. Are you spending 80% of your time on GPA and 20% on building?

Flip it.

Spend 20% protecting your GPA (keep it above 3.5, just enough), and 80% building things that matter.

♻️ Restack if you know somebody who may benefit with this info.

Part 3: Your Resume Isn’t the Problem. Your Artifacts Are.

Parents ask me every week: “Should I use an AI resume builder?”

Most resume AI tools solve layer 2 of a 3-layer problem. And most students never solve layer 1.

Layer 1 (The Real Problem): You don’t have shipped, deployed projects with real URLs.

Layer 2 (Where AI helps): Resume formatting and language polish.

Layer 3 (The Strategy): How you present those artifacts in warm outreach.

AI resume builders are pointless if you skip layer 1. A polished resume describing imaginary projects still gets rejected. A rough resume with a GitHub link to real shipped work gets the interview.

An AI CV tool cannot generate the artifact for you. It cannot build the Bridge Projects. It cannot get your app live. Those require you.

What AI resume builders DO: they make your real artifacts sound better. But only if the artifacts already exist.

The Logic Gate you need to be concerned about. Before running your resume through AI, ask: Do I have two live projects to link to? If no, that’s your first move. Once you have deployed work, then polish the resume.

Part 4: How to Master the Interview Process

August is application season. You’ll get an interview. Most students prepare for all three screens the same way. Most fail in the first minute.

They train for the wrong thing.

Over twenty years hiring at Intel, HP, Toyota, I have watched this. The freeze hits. The panic settles. Most kids have never practiced explaining themselves under pressure. They trained for coding problems. They memorized LeetCode. But they never rehearsed the actual skill that matters: surviving the freeze and communicating like an engineer.

Here’s the sequence that actually works:

Weeks 1-3: Behavioral (Start here)

This is where most students crash and burn. STAR stories. Situation-Task-Action-Result. You need five of them ready. Not memorized but well rehearsed. Tell them out loud. To a mirror. To a friend. Practice until they feel natural.

Behavioral is the foundation. It builds confidence. By week 3, you are not thinking about how to answer anymore. You just answer.

Weeks 2-7: Technical (Overlap is fine)

Blind 75 problems. One per day. Not memorizing solutions but understanding patterns. When you code, explain your thinking out loud. This matters more than perfect syntax. An interviewer wants to hear your brain working, not a silent solution.

Weeks 6-8: System Design

Learn the framework. Do one or two mock problems. You will feel unprepared. That’s normal. The goal isn’t confidence. The goal is having a system to survive the freeze.

You will not feel ready. That’s exactly right.

Why This Sequence Works:

Most students do technical first because coding feels more “real.” But that’s backward. You nail the behavioral round, you build momentum and confidence for technical. You crash behavioral, you are already nervous for technical. And system design? That’s the close. By then you are exhausted or energized depending on what came before.

The Communication Layer:

This is the differentiator from 20 years of hiring. Communication beats coding perfection. The freeze is beatable. But only if you have practiced explaining yourself under pressure. Most students haven’t.

The Logic Gate is this. Can you tell me three STAR stories right now without hesitation? If not, that’s week 1. Start talking, not coding.

Part 5: The One Skill That Decides Your Return Offer

You survived the interview. They offered you an internship. Congratulations. Now forget everything you just learned.

The return offer isn’t about technical brilliance. It’s about one skill, which is making life easy for the hiring manager.

Think from their side. You got assigned a task. Initial onboarding is done. What do they want? They want you to run with it. They want you to ask clarifying questions if something’s ambiguous. They want you to come with suggestions, not just problems. They want to give you direction once, then trust the execution.

This is the gap, the Builder Delta, between listing what you did and functioning independently. Most return offers are won or lost here.

The intern who needs constant check-ins? No return offer. The intern who scopes work, handles constraints, integrates with existing systems, ships something, and learns from real users? Return offer.

From a Hiring Manager’s Perspective:

“This intern doesn’t need managing. They need direction and a check-in, then they run.”

That’s the skill. Not coding speed. Not algorithm knowledge. Independent execution after the initial direction.

The Logic Gate looks like this. When your manager gives you a task, do you come back with questions + suggestions? Or do you disappear and hope it’s right? The first pattern gets the return offer.

Part 6: Your Kid’s Salary Negotiation Fear Is Costing Them Thousands

You got the return offer. Full-time offer coming. Congratulations again.

Now most grads do something terrible: they accept the first number.

Why? Fear. Imposter syndrome. “I should be grateful.” All totally normal. Also totally costly.

A kid who negotiates $8K higher on a starting salary is earning $400K more over a 50-year career (4% raises compound). One negotiation move is needed. One conversation is all you need.

Here’s how you approach it.

Step 1: Get Data

Use levels.fyi. Search by company, job level, location, function. See what others earned at that company. See competing offers.

Step 2: Prepare Your Anchor

Bring 2-3 comparable data points. Not imaginary. Real levels.fyi data. Real competing offers if you have them.

Step 3: Ask the Question

“What’s the negotiation window here? I’ve researched market rates and I’d like to discuss the offer.”

That’s it. You’re not greedy. You’re informed.

The Framework That Works:

Data + Confidence = Proactive stance

(vs. passive acceptance, which is the default)

Most grads don’t negotiate at first offer. It’s not because the offer was perfect. It’s because they didn’t know negotiation was an option.

The Logic Gate is this. Before you accept any offer, run it through levels.fyi. If you’re below market, ask one question: “What’s the negotiation window?” If they say no, you know. If they say yes, you just negotiated.

What’s Next

Get the complete guides on Gumroad:

→ The College CS Toolkit, Full step-by-step guides for all six parts, Bridge Projects templates, interview frameworks, and negotiation scripts.

→ High School CS Internship Playbook, If you are still in high school and need the 9-week acceleration plan to land freshman internships.

→ Project Blue Print, Start your portfolio building in real time as you move through the toolkit.

Start with the Bridge Projects. This week. Not next month. You have got a three-month advantage if you move now.

The kids who start freshman are the ones who compress years into months.

Let’s build.
Masaru

P.S. If your student is sitting on a strong transcript and wondering why internship offers are not coming, reply to this post. We can look at what student has built so far and figure out the fastest next move.

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

Rice Essay Prompts (2026-2027): What a Merit-Winning Answer Sounds Like

The Engineering Dad — Thu, 06 Aug 2026 11:26:23 GMT

You have been reading what I have experienced, and sometimes it helps to listen to others who have gone through the same experience. Ultimately, your child will be writing essays and going through the full college application process, so let’s hear from one of my children who went though the same experience that your child may be going through.

Guesswork

Most applicants aiming for admission in 2027 and onward look at the Rice University essay prompts and rely on guesswork.

They guess at the meaning of the question, “What appeals to you about Rice University?” They guess how to answer the question about community contribution, or guess which version of Prompt 3 they should tackle.

But my daughter didn’t guess. She thought through each prompt carefully, much like an architect solving a problem. What exactly is Rice looking for? What makes for a compelling answer in 150 words? How about 500 words?

As a result, she secured one of the rarest scholarship packages Rice offers.

This video explains exactly how she answered all three essay prompts. It’s not a template or a formula. It breaks down what she actually wrote, why it was effective, and what she would change if she were writing it today.

If your child is aiming for Rice University in this year’s admissions cycle, please watch this video before they start writing their essays. Then, apply what they learn to their own drafts. Good luck!!

RICE ESSAY PROMPTS (2026-2027)

Rice admin page recently updated the upcoming prompts. The prompts are similar to what my children went through.

1. Please explain what draws you to the academic areas you selected above and how you hope to explore them at Rice University. 150 word limit.

2. Based upon your exploration of Rice University, what elements of the Rice experience appeal to you? 150 word limit.

3. Please respond to one of the following prompts to explore how you will contribute to the Rice community:

The Residential College System is at the heart of Rice student life and is heavily influenced by the particular cultural traditions and unique life experiences each student brings. What life experiences and/or unique perspectives are you looking forward to sharing with fellow Owls in the residential college system? 500 word limit.

Rice is strengthened by its diverse community of learning and discovery that produces leaders and change agents across the spectrum of human endeavor. What perspectives shaped by your background, experiences, upbringing, and/or cultural identity inspire you to join our community of change agents at Rice? 500 word limit.

Watch Her Break It Down

She walks through all three prompts. Pay attention to what she reveals and what she leaves out. Note where she delves deep and why. Pay attention to the level of detail.

Then, try applying that same mindset to your own essay drafts.

♻️ Restack if you know of somebody who benefits with this info.

Let’s build.

Masaru
The Engineering Dad

Your Kid's Salary Negotiation Fear Is Costing Them Thousands

The Engineering Dad — Tue, 04 Aug 2026 11:33:25 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

You have reached the last section of the College CS Toolkit preview. By knowing these, you are well ahead of the steep competitions among other students. You do need to take actions based on the info you received, don’t just read, and not taking any actions. Start today.

Nothing is set in stone… yet

When your child receives a job offer email and sees the salary figure, they often assume it is set in stone. Whether the amount meets their expectations or falls short, they tend to conclude, rightly or wrongly, that there is no room for negotiation. Consequently, they accept the offer immediately without attempting to negotiate.

But what is actually happening? The offer sent by the recruiter is merely the initial offer, not the final one. Since no one tells them this, they remain unaware of the reality. As a result, depending on market conditions and the company’s budget, they may miss out on income they could have secured, potentially an extra $2,000 or even over $10,000.

The misconception causing this loss is simple. They fear that if they try to negotiate, the recruiter will rescind the offer and give the job to someone else. In reality, however, that is not the case.

By the time a recruiter extends an offer, they have already spent weeks evaluating your child through a multi-stage interview process involving several team members. This process entails a massive investment of time, coordination effort, and decision-making. Once an offer has been made, the company wants to close the deal rather than start the entire process over by searching for a new candidate.

That is precisely why they anticipate negotiations. As I discussed in Confessions of a Tech Hiring Manager, the hiring process is costly. After extending an offer, the goal is to finalize the hire, not to restart from scratch. Negotiation is considered a normal part of the process.

I have worked in recruitment for twenty years and extended a large number of offers. Yet, I have never once rescinded an offer simply because a candidate requested a salary increase that fell within the market range. While I might adjust the offer amount or explain budgetary constraints, I would never restart the candidate selection process from the beginning. That is because the cost of doing so would be far higher than the few thousand dollars in additional pay.

♻️ Restack if you like what you are reading so far.

How to prepare for negotiation?

So, how do you determine the market rate? Make use of the website like “levels.fyi” Simply enter the company name, job function, job level, and location to search. This allows you to see the terms offered to other students at that company and the salary levels at competing firms. Compare your child’s offer against this data. The principles I outlined in The Builder vs. Consumer Trap apply here as well, combining “data” with “confidence” allows you to shift from a passive stance to proactive action. Tools that let you check benchmarks make this possible.

If the offered amount appears to be a few thousand dollars below the market rate, it is worth negotiating. If the offer is $5,000 or $10,000 below the market rate, you should definitely negotiate. All it takes is sending a concise, straightforward email to the recruiter, backed by market data that supports your counter-offer.

Template email for offer negotiation

“Thank you for the offer. I am very interested in this position. I looked up salary data on ‘levels.fyi’ for the [Job Title] role at [Company Name] in [Location], and the market rate is typically between $[X] and $[Y]. The current offer is $[Z], is there any room for the team to adjust this closer to [Desired Amount]?”

Many recruiters will either say “yes” on the spot or ask for a few days to check with the department handling the budget. Either way, you establish a record and evidence that you negotiated within the market range. This type of action carries significant weight. It demonstrates that you respect yourself and have conducted thorough research beforehand. If the salary amount can’t be changed, you can ask for other benefits like more PTO, remote/hybrid working days, etc. So don’t just take their salary number is the only thing to negotiate.

Would negotiating be risky?

Your child might think that negotiating risks losing the job offer. In reality, however, by not negotiating, they risk losing thousands of dollars. Fear comes at a higher cost than negotiation does.

If your student is sitting on a strong transcript right now and wondering why the offers are not coming, feel free to reply to this post. We can look at what they have built so far and figure out the fastest next move.

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade.

♻️ Restack if you know your friend who may benefit with this info.

Subscribe now

She Turned Down Caltech and Columbia for Rice. Here's What Decided the Offer.

The Engineering Dad — Thu, 30 Jul 2026 11:35:06 GMT

Last week in a Class of 2027 parents group, someone asked whether Rice engineering was cut-throat like CMU or a little more balanced. My daughter graduated from Rice CS a couple of years ago with a full-time SWE offer at graduation. She had turned down Caltech and Columbia to get there.

Honestly, for us parents, money was first, everything else was second. So full ride was attractive. However, we looked into academics and her future goals. She wanted to become a SWE and work as quick as possible to gain work experience. She had concerns on Caltech’s 5 physics course requirements that each student has to take. Caltech was the most expensive of all even with FA. She didn’t want to become a professor, in which case she might have picked Caltech.

But as a Tech Dad who saw my youngest choose Rice over Caltech and Columbia, I know there is a hidden variable in the Source Code of these schools that most parents miss.

, Comparing the Caltech Grind to the Georgia Tech and Rice Efficiency Engines

We ran the Net Price Calculator on every school before the first campus visit. Rice is a high reach or lottery for most applicants, and the chance of merit aid for CS is slim.

Most parents compare Net Price (which is fake) instead of The Gap (which is real).

, The Apples-to-Apples Calculator

While my child enjoyed Rice very much and came to the same outcome as becoming a SWE, if we had to pay the full COA, we probably would have changed the decision.

For Texas residents and families with strong in-state options, UC Berkeley, UIUC, Georgia Tech, and Purdue produce comparable CS and engineering outcomes at a lower out-of-pocket cost.

As I covered in The $20,000 Rumor, the merit numbers circulating in Facebook groups are often off, sometimes by $20,000 or more.

What the Rice Name Actually Does

Some people call Rice the Harvard of the West, and its general ranking at top 20 in US News shows it is well known in general. Their residential college system is famous and the environment is collaborative. Because of this, Rice name itself goes comparable with other top tier schools. Additional advantage is that if a student comes from top 5 CS schools for instance, people think you must be smart and can deliver. Not all the time is the reality, then that will actually negatively damage the school’s reputation. Rice name might be considered second tier in engineering, but student body is as compatitive and smart as top tier school ones. In that regard, students can perform well in a rather relaxed perspective.

Once you are in the Top 50 of ABET-accredited engineering programs, the ranking number is noise. As a hiring manager, I do not care if a candidate went to the #7 school or the #12 school.

, The CS Showdown

After that, it’s the builder record that closes the offer.

What Career Services Won’t Say Out Loud

Most parents don’t ask about on-campus recruiting until their student is already enrolled. Twenty years hiring engineers at Intel, HP, and Toyota taught me what the recruiting reality looks like.

This is a great opportunity for students to realize they have to go above and beyond. They will have to think of how to differentiate themselves from others, but Rice has resources to enable that.

The companies that recruit on campus lean toward the energy sector. If your student’s goal is Amazon, Google, or Microsoft, they’re sending their own applications and prepping their own technical interviews without a recruiter coming to them.

What She’d Already Built Before She Got There

The offer she landed at graduation traced back to Boone County, Kentucky. She was one of the founding members of the county maker space, built through the FTC robotics team I coached there, along with other FLL and FTC groups in the area.

Maker space allowed her to come up with what items to buy that students build. She had to also do test builds so she could decide and influence what items would be a good learning tool. Ultimately, this maker space would be used by the whole county’s students, middle and high school, which passed my builder framework level 3: build things that are used by others.

As I covered in The Builder vs. Consumer Framework, builders who create things used by others are what selective programs are selecting for, and what hiring managers remember.

For a hiring manager, a maker space serving the whole county’s students weighs differently than a project built for a grade.

The AP vs Dual Enrollment Question

In a separate thread that same week, a parent in the College Admissions Experts group was asking which was better for a rising senior targeting Rice and Caltech for MechE.

For MechE in particular, AP Calc BC is the better move. Selective admissions officers view AP as a trusted, standardized benchmark and many colleges have published credit transfer lists and the process is quick (Rice has this). Rice accepts AP credit straightforwardly (with a 4 or 5), whereas DE credits have to go through a much stricter departmental evaluation. Choose AP Calc BC. It carries more weight in admissions and transfers far more reliably.

By junior year, the strongest Rice University admissions candidates for engineering have AP Calc BC done, with Multivariable Calc or AP Physics C on deck for senior year.

Next Week. The Essay That Got Her In.

If you’re targeting Rice, the essay is where most families underestimate the work.

Rice recently released their essay prompt for 2027, which are similar to what my child wrote. Rice box essay is unique and many parents don’t have good ideas on what to include. That also provides great ideas and overviews on the rice prompts. The goal is to make sure parents are as familiar in the application essay prompts.

Next week on the main site, I’m sharing the video where she reads the essay herself. Watch it before the application window opens.

Let’s build.

Masaru

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade.

Support the Workshop: I spend my spare cycles auditing the Source Code of admissions so you do not have to. If you have found value in these logic gates, you can buy me a coffee to help keep this workshop running and the guides open-source.

The One Skill That Decides Your Return Offer

The Engineering Dad — Tue, 28 Jul 2026 11:11:06 GMT

The College CS Toolkit is live on Gumroad. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

A quick look at how internship goes

HR’s onboarding covers badge pickup, laptop setup, and the parking garage. The first day is structured. None of that tells your student what the manager is watching for starting in week two.

I don’t plan on babysitting you, just don’t have time. That is the reality from the hiring manager’s side, and we expect the person to go on his own with minimal guidance.

I have watched this play out over twenty years at Intel, HP, Toyota and the top-tier Children’s hospital. Three types of interns show up every summer. Two of them don’t get a return offer. The difference rarely comes down to technical skill. As I covered in Confessions of a Tech Hiring Manager: Why I Ignore 4.0 GPAs, the hiring manager isn’t measuring credentials. They are measuring whether you can function without being led.

The Needy Intern

Needy intern is just a distraction for the other side. It may be ok on the first few days while onboarding, but student should improve drastically just within days.

By week two, it’s a pattern. A manager fielding question after question from one intern starts making quiet calculations about whether that seat makes sense next summer.

“In today’s highly competitive market, companies do not hire students simply because they want to learn. They hire students because they have demonstrated the ability to build.”

Why “Looking for an Internship” is the Worst Way to Find One

If you look around FB groups, parents frequently keep asking, “I am looking for internship for my child in this area.”. That’s really a wrong question to be asking.

The Silent Intern

When silent, I can’t tell where you are in the process. Some may be ok and just quiet, while others may be having a hard time, but just don’t have the courage to ask around. You must speak up when needing help.

Question Calibration is the hardest skill in a first internship. Knowing how long to try on your own, and what to bring when you do ask, is something most students figure out too late. Part 5 of the College CS Toolkit covers exactly that threshold with a concrete decision framework.

College CS Toolkit is live. It cover all these college internship and hiring secrets. Knowing these will put you ahead of others. Substack readers get 25% off with code COLLEGE25 at checkout. Grab your copy here. After the initial sale, the price will go back to $59.

Make Life Easy for the Hiring Manager

Think in terms of the hiring manager side. You have assigned a task to this intern, and the initial onboarding is done. What do you expect? You want the intern to do his own thing based on the task requirements. If there are ambiguity, you want the intern to ask you and come up with suggestions so you can guide them. Make life easy for the hiring manager side. That’s the key.

This intern doesn’t need managing. They need a direction and a check-in, and they run from there. As I covered in The Builder Delta: Why “Club President” Isn’t an Artifact, the gap between listing what you’ve done and demonstrating you can function independently is where most return offers are won or lost.

♻️ Share if you like what you are reading so far.

Culture Fit Closes the Return Offer

Two interns can have the same GitHub profile, nearly identical test results, and similar technical output over ten weeks. The one who gets the return offer is usually the one the team wants to sit next to again.

The one that’s easy to work with is somebody who doesn’t have egos (that he knows everything), and he is willing to listen to what we say for guidance. He is social and friendly, approachable. He comes to work physically and visible. Comes to intern get together with others and genuinely try to find commonality among others and see if that is the workplace he wants to be in and contribute to the overall goal of the organization. Technical knowledge is important, but that is trainable, and learnable. But socialization skills are something that’s hard to change or learn.

As I covered in Building Culture: Why I Spent More Time on the Floor Than the Roof, being physically present and visible in the workspace isn’t optional. It’s how trust gets built on a team. btw, if you haven’t read this article before, take a look, as I am proud of the progress made here for my DIY work (not college admission specific).

The 90-Day Map

Part 5 of the College CS Toolkit covers the week-by-week sequence, the Question Calibration framework, and what to say in the final two weeks before anyone asks about a return offer. Most interns don’t know the decision often happens before anyone mentions it. As I wrote in The Builder vs. Consumer Framework, what separates a builder from a consumer shows up most clearly when direction is scarce, and the internship is the first real test of that.

Finally, and most importantly, it is possible you may get multiple offers, and ideally, you want to be able to negotiate to get the best one. But how do you know which one is the best one? The College CS Toolkit gives you a glimpse of what to look for in the offer letter so you can fairly compare. That’s a skill college career office may not provide enough guidance and that’s a critical skill you need because that could change your financial future landscape.

If your student’s internship starts in the next 60 days, this is the guide to send them the night before day one.

Get the full College CS Toolkit here.

These interactions require some practice, but practice makes perfect. It is doable and many students do get internship while in college and return offers.

You can do this.

Let’s build.
Masaru

You can grab the High School CS Toolkit here.

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade.

Pangram Flagged Me as 100% AI. It hurt my feelings.

The Engineering Dad — Sun, 26 Jul 2026 20:45:42 GMT

This is a special Sunday edition because there is a change that happened in substack that is causing substack writers to be possibly concerned about. I want to let you know this so in case you encounter this in future substack articles, you know why.

What is Pangram?

Substack started using Pangram AI Scanner this past week. As a native Japanese speaker with a long interest in language mechanics, I naturally wanted to try Pangram on my own writing.

I ran several past Substack articles through this detection tool, and the results were not what I hoped I would see, but somewhat expected. Some articles were 100% classified as “AI-generated,” while others were 100% classified as “human-generated.” There was no middle ground. There were no partial scores, no nuances, no intermediate areas. Even this binary result alone revealed a fundamental flaw in the detection tool’s mechanism.

I needed to understand what Pangram was actually measuring. In 2013, in an NLP class at Columbia University, I built a system for evaluating classification models. We split samples, trained the model, tested it with the remaining samples, and finally verified it against human-verified ground truth data (golden copies). This process allowed us to calculate a percentage of accuracy and measure whether our model was better than the professor’s. I still remember that experience vividly. If a classification tool only produces 0% or 100% results, there is a problem with the evaluation system itself. True accuracy should exhibit some kind of distribution within that range. But Pangram does not for my articles scanned. It makes judgments based on a strict threshold. Either/or. No middle ground. Like a malfunctioning detector. As it turns out, others on substack appear to be experiencing something different, some have partial AI and human scores. I wanted to learn more.

What Pangram measures

So, what exactly is Pangram measuring?

Pangram does not measure “authenticity” (whether it was written by a human) at all. This is the fundamental problem. This detection tool measures “perplexity,” an indicator of how predictable a word’s selection is when compared to patterns in the training data. Authenticity and perplexity are not the same thing. They are not even remotely similar. Nevertheless, Pangram treats high perplexity as if it were proof that something “was not written by a human.” This single concept explains all the mistakes this detection tool makes regarding authenticity.

I recall a specific example from Columbia University that illustrates this point perfectly. I was working with the three-word trigram “I were” from the phrase “I wish I were an angel.” This is grammatically correct. The subjunctive mood is required after “wish.” However, since most people write “I was” instead, “I were” is rare in typical writing. When a classification tool recognizes “I were,” it registers it as “rare.” “Rare” is flagged as “suspicious.” Based on the detection threshold, “rare” is determined to be “probably AI-generated.” However, the writer of that sentence that Pangram determined to be AI-generated was, in fact, me.

Here is something I discovered while researching Pangram’s actual behavior. I took one of my articles that Pangram had flagged as 100% AI. I translated it to Japanese using Google Translate, then translated the Japanese version back to English. The result was grammatically awkward. The phrasing seemed strange. A native English reader would have noticed something was off (the sentence structures did not match typical English patterns). However, when I ran that double-translated version through Pangram, it scored as 100% human. Same content. Same ideas. Different language patterns. Completely different result. This tells me Pangram is not measuring whether text is “authentic.” It is measuring whether the word patterns match what its training data learned as “English AI patterns.” Machine translation produces different “perplexity” patterns (statistical word predictability) because the words go through a different language model. Pangram has no training data (or minimal training data as such) on translated text, so it does not flag it. This single test proves the detector is measuring the wrong variable entirely.

♻️ Restack if you like what you are reading so far.

People learn English in different ways

Consider how I learned English. I started learning in the fourth grade of elementary school, attended a private school with six hours of English classes per week, and went on to an American university in Japan. There, I acquired a formal and orderly way of constructing specific sentences. In formal writing, contractions are avoided, and correct subjunctive forms are used. Because of this training, my thinking is based on standard English grammar. As a result, compared to casual English speakers, statistically unusual word combinations are produced. When Pangram analyzes my writing, it detects these unusual grammatical structures and flags them as “heterogeneous.” This is because, in its criteria, “heterogeneous” equals “AI.” However, the writer that Pangram flagged is still me.

This can happen to any ESL (English as Second Language, non-native English speaker) writer who has received formal grammar education. In Japanese conversation, context and hierarchical relationships are emphasized. English, on the other hand, is much more flexible, and language use does not change depending on the speaker’s social status. However, when Japanese ESL learners bring their formal construction methods into English writing, Pangram recognizes word patterns that it has learned to associate with statistically “heterogeneous” (unusual) patterns. These rare patterns are then judged by Pangram as being from AI. Nevertheless, these ESL writers possess a genuine “authenticity” inherent to their writing style. It definitely exists. It is simply that they have been trained in a way that differs from the patterns assumed by Pangram’s training data.

This is not judging the “writer’s individuality.” It is merely measuring statistical “rarity” and assuming it is proof that it was written by AI.

What do I do on hiring?

Here is what I actually prioritize when hiring. My 20 years of experience at Intel, HP, Toyota, and top-tier children’s hospitals have taught me what is important in terms of “authenticity” (the genuine expression of capability and character). I seek interns with the fundamental technical knowledge relevant to their role. I want them to be approachable and easy to work with. Most importantly, they must be able to work autonomously without constant supervision. They deliver results, and they explain the challenges they faced and how they overcame them. That is what “authenticity” in practice is all about.

I have never evaluated someone based on how predictable their word choices are. I do not care whether they use unusual grammatical structures. What I value is whether they delivered results, clearly communicated the problems they faced, and demonstrated autonomous judgment. Grammatical predictability is unrelated to authenticity. If Pangram had flagged one of my interns, who had returned to work with me, as “100% AI” simply because their writing had a well-structured format, then that detection tool would be making a completely wrong judgment about themselves and what they produced. However, because Pangram does not understand authenticity, thousands of ESL writers are currently experiencing this very problem. The detection tool is simply measuring the wrong qualities and labeling them as “truth.”

When you see a Pangram warning, you should not ask about grammar. You should ask about the work itself. Does the writing solve a real problem? Are the ideas concrete or cliché? Does the writer explain how they overcame real obstacles? Can you see that they acted autonomously or constantly sought help? These are the true signals, the true indicators of authenticity.

When reading someone’s writing, pay attention to these aspects. Look for specificity and evidence of real-life experience. Find moments where the writer speaks of problems they have personally faced and solved. That is where authenticity lies. Punctuation, contractions, or Pangram’s “perplexity” score are irrelevant.

Going forward

I am not going to change my writing style to pass detection tools that measure wrong. I recommend the same to all ESL writers dealing with Pangram. Understand what the detection tools are measuring. Be skeptical of tools that label perfect grammar as “evidence of a lack of authenticity.” Detectors that measure “perplexity” and detectors that measure “authenticity” are entirely different. Recognize this fact. Authenticity cannot be measured by Pangram or other AI detection tools. It can only be recognized by reading actual works and experiencing the writer’s thought process, genuine experiences, and truly independent judgment that are evident throughout. That is the “signal” that a true connoisseur should value.

As I covered in Confessions of a Tech Hiring Manager: Why I Ignore Perfect Resumes, authenticity shows up in what someone builds and how they explain the work, not in grammar predictability.

What are your thoughts on AI scanners? Do you discredit somebody writing if an AI scanner says it’s AI generated in this AI widespread technological advanced era we are currently in? What do you look for in any writing?
I want to hear your opinions, comment below.

If your child is in college, and want to know secrets of how to get college level internships and ultimately a full-time return offer, there are frameworks you can follow specifically for builders, not typical syllabus consumers. Know this framework early on will change your job prospects. Check here to be in-the-know.

Share the post if other parents and students you know may benefit with this info.

The #34 School Might Be Your Best Engineering Bet

The Engineering Dad — Thu, 23 Jul 2026 11:57:05 GMT

The College CS Toolkit is live as of this morning on Gumroad. If you’re on my current Gumroad email list, check your inbox for the exclusive early access code, 50% off the first 5 copies at $49. Everything you need from what to build, how to pass resume test through first 90 days on the job. Grab your copy here!
♻️ Forward to your fellow parents with college students!

I am a Tech dad of 3CS grads (Cornell, and Rice chosen over Caltech/Columbia). Rising seniors are now starting to seriously looking to construct a list of colleges to apply in the fall. Read this before you decide on the final list.

The school ranked 34th on the national university list might be the strongest aerospace engineering program in your region. No junior year prep guide will flag that for you, because no junior year prep guide knows what your student’s major is.

Every family with a rising junior or senior has seen a version of the same checklist. Look around in Facebook feed. College counselors and education centers tell you exactly the same thing. Build your story. Visit schools. Create your college list. Start looking for money. Set up your recommendation letters. All of that is correct. None of it is enough for a STEM family in junior year. I see parents on FB groups asking “oh, we are visiting Ivys, here and there, what should we ask?”. Statistically, your child has more chances to get rejected than accepted. Why would you go visit before your child gets accepted, but your child loves the campus so much that it is now his or her dream school. If statistics hold (meaning rejection), that can be a mental disaster for the child. I and you both want to prevent this situation as much as possible.

Most junior year prep guides assume the right ranking source is obvious. For a student targeting CS, EE, aerospace, or biomedical engineering, it is not as obvious, however. A school ranked 8th nationally might have a mediocre electrical engineering department. The school ranked 29th might place the majority of its graduates into the labs and companies your student wants to reach. Generic prep guides treat college selection as a school-level decision. For STEM families, it is a program-level decision, and junior year is when the mental model gets set.

I spent 20 years hiring engineers across Intel, HP, Toyota and one of the top Children’s hospital in the nation. I coached competitive robotics from FLL through world-level FTC. Three of my kids went through CS and engineering admissions at Cornell, Rice, and Northeastern. I have been on both sides, reviewing transcripts as a hiring manager and building college lists as a father. Most college counselors work from school-level reputation data, which is the right tool for general admissions. It is the wrong tool for choosing between two aerospace departments. That distinction is what this article is about.

♻️ Restack if you like what you are reading so far.

The ranking source the checklist skips

Standard advice says to research academics, outcomes, culture, and costs before building your list. It does not name a data source.

For CS, CSRankings.org is the right source. It measures faculty publication output at the most selective research conferences using DBLP data, a computer science bibliography that is harder to game than a reputation survey. More importantly, you can filter by subfield. A school that ranks 15th overall in CS might rank 4th in computer systems or 7th in human-computer interaction. If your student wants to work in AI infrastructure or machine learning research, the overall CS rank is the wrong number. The subfield rank is what matters.

Once you are in the Top 50 of ABET-accredited engineering programs, the ranking number is noise.

The CS Showdown: How to Choose Between Top Engineering Schools

What replaces the ranking number is research output, program flavor, and the pipeline a department has built into specific industries. CSRankings surfaces that. US News does not.

For EE and aerospace, the Shanghai Global Ranking of Academic Subjects is the equivalent source. It uses research output and citation data across 57 subjects rather than reputation surveys. A school your guide lists as a match might be a real reach in EE. A school you have not considered might be a recognized leader in aeronautical engineering by the researchers who work in the field.

For biomedical informatics, the NIH maintains a hard list of T15 training grants, federal programs awarded to programs that passed peer-reviewed scientific standards. No marketing budget manufactures a T15 grant. No reputation survey replicates it. If a school holds one, it cleared federal scientific review, and that fact will not appear in any junior year prep checklist.

The money decision hiding in junior year

Standard checklists tell families to start looking for scholarships, local organizations, community foundations, employer programs. That advice is fine as far as it goes.

The real money decision in junior year is running the Net Price Calculator before you get emotionally attached to a school. At most top-ranked CS programs and out-of-state schools with selective engineering departments, automatic merit aid is near zero. MIT, Carnegie Mellon, Georgia Tech out of state, Purdue out of state, the sticker price and the real price sit close together. Parent groups on Facebook routinely get this number wrong. As I wrote in The $20,000 Rumor:

My own child received a $25,000 merit award at Northeastern (NEU) and other parents in the thread were citing $15k and $22k tiers.

The number you read in a thread is almost never the right number. The NPC is the one you must watch.

You cannot trust the bottom-line number printed on the financial aid letter.

The Apples-to-Apples Calculator: Stop Guessing and Start Comparing

Run the NPC in junior year, before the campus visit, not after. Once your student has walked the campus and met the robotics team, the number feels like a problem to solve rather than data to act on. As the Trojan Horse in Your Financial Aid Letter breaks down, the way aid packages are structured is designed to obscure the real price, not clarify it.

A financial safety school and an academic safety school are different categories, and this distinction does not appear in standard junior year prep. A school that is an academic sure thing might still cost $65,000 per year out of pocket. A real financial safety school is one where you know the merit threshold before you apply, that threshold is achievable for your student, and the award is automatic rather than competitive. Junior year is when to identify those schools, before the emotional weight of the list makes the math hard to act on.

Two decisions that look like preferences

The ranking and money sections above are about building the right list. Two items from the standard checklist affect what happens after the list is built.

The first is Early Action (EA). Most acceptances at selective schools happen in the EA round. A list deliberately weighted toward EA-available schools is not an aggressive strategy. It is an accurate one, because it is weighted toward where the probability actually sits. A list built around campus culture and visit impressions, without thinking through EA distribution, loses acceptance probability without knowing it.

The second is co-op access. Standard prep guides mention Northeastern’s co-op program and treat it roughly the same as summer internship programs at other schools. They are not the same.

A full-term co-op produces a student who graduates with 12 to 18 months of professional experience. A summer internship program produces a student with two or three summer stints. In a job market where entry-level resumes get reviewed in under 30 seconds, those months of difference separate candidates from the same graduation year.

The gap between listing experience and demonstrating it is where most applications die.

The Builder Delta: Why Club President Isn’t an Artifact

A co-op builds the demonstration. A summer stint builds the line on a resume. Choosing between a co-op school and an internship school looks like a campus culture preference. Most STEM families in junior year make that choice without knowing what they are deciding.

Three additions for the STEM junior year checklist

The standard junior year prep guide is correct in its categories. It is written for everyone, which means it fits no specific STEM major. Three additions close the gap.

Before you build the list, choose the right ranking source for your student’s major. CSRankings for CS, with subfield filters applied. Shanghai GRAS for EE and aerospace. NIH T15 training grants for biomedical informatics. QuantNet for quantitative finance. Use program-level data as the primary filter. Add university-level rankings as a secondary check for schools that clear the program threshold.

Before the campus visit, run the Net Price Calculator on any school your student has gotten attached to. If the number does not work for your budget, the visit is not worth scheduling. This is not pessimism about the school. It is the decision sequence that keeps the STEM college list usable through senior year.

Before you build the list, settle the co-op question. If your student needs 12-plus months of professional experience to be competitive in their target field, build toward full-term co-op programs. If summer internships are sufficient, the pool of strong options expands. This is a major-specific answer that should be made before the first campus tour.

None of this requires a different checklist format. It requires knowing that STEM junior year college prep is a program-level research problem, not a school-prestige comparison.

My three kids navigated this process at Cornell, Rice, and Northeastern. I did not get all of it right the first time. By the third round I understood what the standard checklist was missing and why it was hard to find that information in one place.

Let’s build.

Masaru

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade.

How to Master the Interview Process

The Engineering Dad — Tue, 21 Jul 2026 11:51:08 GMT

August is application season. Your student will get an internship interview. They will face three screens. Most kids prepare for all three the same way. Most fail within the first minute. They trained for the wrong thing.

I’ve watched this for twenty years across Intel, HP, and Toyota. I’ve hired interns and trained them. I’ve coached robotics teams to world championships where communication was the difference between a medal and a DNF. The interview is not one event. It’s three events. Different skill, different prep, different timeline for each.

Here’s what they actually face.

The Behavioral Interview. STAR Method.

Your student walks in the room. The interviewer asks a question.

Tell me about a time you failed.

Tell me about a time you disagreed with someone.

Tell me about a time you had to learn something new under pressure.

These are not personality questions. They’re not asking if your kid is nice. They’re testing whether your student can structure their thinking. Can they tell a coherent story under stress. Can they show judgment.

Most students ramble. They jump to details. They miss the point. The interviewer stops listening at minute two.

This is where the STAR method lives. Every behavioral question has the same structure.

S. Situation. Set the scene, briefly. Not a fifteen-minute story. One sentence.

T. Task. What was your specific responsibility.

A. Action. What did you personally do. Not the team. Not the manager. You.

R. Result. What happened. Use numbers if you can. This is the payoff.

Your student has not thought through real stories yet. That’s the gap. It’s not intelligence. It’s preparation.

As I covered in Why Looking for an Internship is the Worst Way to Find One, companies do not hire students simply because they want to learn. They hire students because they have demonstrated the ability to build. The STAR method makes that demonstration audible.

The best students prepare three stories. Real ones. One for conflict, one for failure, one for leadership. That’s enough. Adapt those three to any behavioral question the interviewer throws at them. The story changes slightly. The structure stays.

This is not a secret. Amazon, Google, Microsoft explicitly teach this. But most high school students have never heard of STAR. They walk in and hope. Hope doesn’t work.

Timeline: Three weeks to prepare behavioral. Story prep happens first. It’s easier than technical. It builds confidence. Start here.

The Technical Screen. Blind 75. Pattern Recognition.

After behavioral comes the whiteboard. The interviewer hands your student a problem.

Write a function. Given an array of integers, find two that sum to a target.

Your student has never seen this exact problem. They panic. They know algorithms exist. They don’t know which one.

This is where LeetCode lives. Your student has not solved 2,000 problems. Neither did the interviewer when they interviewed. But your student needs to recognize patterns.

The Blind 75 is a curated list of 75 essential problems. Arrays, Strings, Trees. These 75 problems cover the underlying patterns that show up in 90 percent of technical interviews. Your student does not need to solve 2,000. They need to understand 75 and be able to adapt.

This takes weeks. Four to six weeks if they focus. One problem per day, with time to understand the pattern, not just memorize the answer.

Here’s the trap. Most students grind problems silently. They get the right answer in their IDE. They think they’re ready. Then they sit in front of a whiteboard and their brain freezes. They can’t explain what they’re doing. The interviewer cannot see their thinking.

The actual skill is verbal problem-solving. Your student needs to talk through their logic. Walk the interviewer through the pattern. Handle the push-back. This is a speaking skill, not a coding skill.

The secret weapon is Big O notation. Most students can solve problems. Few can explain why their solution is O(n) instead of O(n²). That explanation is where you separate yourself. It’s not about being smarter. It’s about thinking clearly under pressure.

Timeline: Four to six weeks to prepare technical. This is the long pole. Start after behavioral stories are solid.

The System Design Lite Screen. The 45-Minute Framework.

If your student made it this far, they get one more screen. System design. Most companies tell interns they won’t get this. But Google, Meta, Stripe increasingly ask. Your student should know the framework.

This is not architecting Netflix. This is a forty-five-minute conversation. Design a URL shortener. Design a notification system. Design a ride-sharing app.

Here’s where most students lose their nerve. The final screen isn’t about coding. It’s about thinking like an engineer.

Most students panic. They don’t know where to start. The framework solves this.

Clarify the scale first. Don’t draw boxes. Ask questions. How many users. What’s the read-to-write ratio. What’s the latency requirement. Most students skip this and start drawing. That’s wrong. You must ask initial questions. That’s how the conversation starts.

Define the API. Then the schema. Visually explain the flow. UI sends data to backend. Backend sanitizes and validates. Backend calls an external API. Response flows back. This proves you understand systems integration.

Discuss trade-offs. PostgreSQL or NoSQL. Cache or load balancing. Why would you pick one over the other. This is where thinking like an engineer shows up.

The whole thing takes forty-five minutes. Your student does not need to be perfect. They need to show they can think out loud and reason about complexity.

Timeline: One to two weeks to understand system design. Not heavy prep. Just know the framework. Do this after technical prep is solid.

How to Sequence the Prep. Six to Eight Weeks.

Weeks 1-3: Behavioral. STAR stories. Practice telling them. Get comfortable.

Weeks 2-7: Technical. Blind 75 problems. One per day. Verbal walkthrough. Understand patterns.

Weeks 6-8: System design. Learn the framework. Do one or two mock problems.

Overlap happens. That’s fine. But start with behavioral. It builds momentum and confidence.

Your student will not feel ready. That’s normal. The goal is not confidence. The goal is having a system to survive the freeze.

The Reality.

Getting the interview is 10 percent. Nailing it is a system. That system is learnable. Your student doesn’t need a magical IQ. They need preparation and the right framework.

Everything you need is in the complete College CS Internship Playbook. Part 4 is all interview mastery.

The playbook covers Parts 1-6 (total 100+ pages), from identity shift to first ninety days on the job.

As I covered in The Builder vs. Consumer Framework, the students who get offers are the ones who moved from consuming credentials to building artifacts before the deadline forced the question. Interviews are no different. You can consume interview tips online. Or you can build your three STAR stories, solve your Blind 75 problems, and own your interview performance.

Let’s build,

Masaru
The Engineering Dad

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade.

Ninety Seconds Inside Your Student's GitHub

The Engineering Dad — Sat, 18 Jul 2026 13:28:12 GMT

Why a Saturday edition?

This is a Saturday edition for you to be in-the-know. I’m sending this as a Saturday special because the 2027 internship application season is about to open, and most students are not ready. Every August, when school starts back, companies begin accepting applications for summer internship positions. GitHub portfolios are reviewed in the first pass. Most student portfolios fail the ninety-second test. If your student wants an internship slot, their evidence needs to be visible before applications close, which happens faster than parents realize. This article exists because you have maybe four weeks to fix it. This is applicable for both high school and college internships.

90 seconds rule

What happens in the next ninety seconds when a reviewer opens your student’s GitHub, I have lived it hundreds of times as a hiring engineer. The engineer clicks the link. She sees two commits. One says final version and one says upload. The README says my project. The three nights your student spent hunting a sensor bug leave no trace. No explanation of why they chose this design over the obvious one. No test results. She closes the tab.

The project was real. The evidence was missing.

When I was reviewing college applications, I had about 100 resumes sitting on my desk from Target Schools. Almost all of them listed President of X Club or Member of Y Society. Here’s what I learned: the resume doesn’t tell me what the student actually built. It tells me what they listed.

Your student probably built something good. Most high school students don’t build anything at all. But when a reviewer opens the GitHub, they can’t tell the difference between a project your kid spent weeks debugging and one that was assembled in a weekend. That’s not a flaw in your student. That’s a visibility problem.

As I’ve written before, the gap between a resume line and the actual work is where most applications die. Your student needs to show the thinking behind the work, not just list that the work exists.

Science fair judges run the same ninety-second evaluation in person. What broke. How did you debug it. Why did you build it this way. The students who win built projects where the thinking is visible.

This is where most students fail. The project is real. The evidence is missing.

What Reviewers Actually See

When I review a student project, I’m looking at four things. Not whether the idea is original. Not whether the code is elegant. I’m looking at whether the student can show their thinking.

Code Quality. Did they write code that works or code that they understand. A single function with no comments and five nested loops is a red flag. Five functions with clear names and one level of nesting is a signal.

Git Workflow. Did they iterate and debug, or did they dump the final version and hope. Two commits named final version and upload tell me the student doesn’t know what version control is for. Twenty commits with a narrative of what broke and how you fixed it tells me the student thinks like an engineer. I’ve seen a student take a single commit and refactor it into twenty honest ones in an afternoon. Same project. Completely different signal.

Troubleshooting Narrative. Did the project work on the first try. If so, that’s fine, mention it. But most projects have a moment. A sensor that doesn’t work. A calculation that’s off. A performance problem that took three days to solve. That moment is gold. That’s where the thinking shows.

Documentation. Can someone else run this. Can they understand what it does and why. The best code is clear. The second best is code with a README that explains what the student was trying to do and why they made the choices they made.

Most students have the thinking. They just don’t show it.

Here’s the shortcut

If your student has built something and it’s sitting on GitHub without a clear narrative, this is the gap. Reviewers spend ninety seconds and move on. Most students don’t get a second look because the first look shows nothing.

I review student projects the way I have reviewed engineering candidates for twenty years at Intel, HP, Toyota and one of the top Children’s hospital in the nation. A scored project audit tells your student exactly which of the four areas needs work, in what order, and how to show the progress. The audit is $199 one time payment, delivered within five business days.

Get the audit here.

Or read on for how to do this yourself.

How to Show Your Thinking

If you want to fix this without the audit, here’s the sequence.

Start with git history. Did you really make twenty commits or did you squash them all into two at the end. If you squashed them, write them out. Not perfectly, not a work of art. Just honest. What did you try. What broke. What did you learn. That narrative is what counts.

Add a troubleshooting section to your README. Not a fancy doc. Just a list. This sensor didn’t work at first because X. I solved it by Y. This taught me Z. Science fair judges eat this up. Reviewers see that the student is thinking, not just coding.

Write comments that explain decisions, not syntax. Not every line. Just the moments where someone else might go a different direction. If you chose this library over that one, say so. If you spent two days on optimization, say so. You’re not boasting. You’re showing your work.

With summer ending and the school year ahead, now is the moment to transition from passive consumption of tutorials to active building with intention. These take a weekend. Maybe two if your git history is really buried. And they change what a reviewer sees.

The other path is the audit. Five business days. A scored report that tells you exactly what to prioritize. Before and after examples so you know what good looks like. Most students complete the fixes in two to three weekends after getting the report.

Pick your path

Here’s what I know after twenty years. The Engineer’s Approach works better, costs less, and actually prepares your child for the job, not just the application. Whether your student fixes their GitHub alone or with the audit, the difference is time and confidence.

Your student built something real. The question is whether the evidence survives the ninety-second test. If you want to know for sure, get the audit.

Get the audit.

Every dollar credits toward mentorship if your student wants ongoing help afterward.

Let’s build,

Masaru
The Engineering Dad

Your Kid's Internship Won't Guarantee a Job Offer (And That's Okay)

Thu, 16 Jul 2026 11:04:18 GMT

The Confusion Parents Have

The truth about college internships and job offers. Here’s what I hear from parents all the time: “Does an internship lead to a job?” “Should my kid take an unpaid one?” “If they intern at Google, are they basically hired?” The answers are messier than LinkedIn would have you believe. Let me cut through it.

I’ve been on both sides of this. I was the kid nervously cold-emailing recruiters for my first internship. I have been the hiring manager reviewing whether to extend a return offer. The view from each side is completely different. And that’s exactly what makes the timing of internships so critical for parents to understand.

What Actually Happens During an Internship

Here’s the hard truth. Most internships aren’t luxury like you may believe. Your kid isn’t shipping the core product feature. They’re not leading the architectural redesign. They’re not even necessarily writing code that makes it to production.

I was watching “The Intern Experience” at a software company recently and it nailed this. One intern was debugging UI issues on mobile. Another was writing test cases. Another was working on documentation. These are real Roblox interns at a major gaming company. Their work looked like this. Not glamorous but useful.

That’s actually the right internship.

As an engineer working alongside interns, I saw this pattern. The intern gets assigned a concrete problem. It’s something the team needs done but it’s not blocking anything critical. It has clear scope. The intern pairs with a senior engineer for the first week, then goes relatively independent. They ship their work or it gets reviewed and shipped. They move to the next problem.

This is different from “busy work internships” I’ve seen at smaller companies where the intern shadows engineers or attends meetings. And it’s different from “research project internships” where the kid spends 12 weeks on something that gets thrown away.

The best internships are the ones where the kid does real work, completes it, and can point to it after and say, “I built that.” Even if “that” is a mobile UI fix or a testing framework. It’s a real accomplishment that somebody could see on github or in production.

Real internship work means solving complete problems: scoping the work, dealing with constraints, integrating with existing systems, shipping, and learning from real users. Students who’ve only written code miss all of that. Stop “Learning to Code” and Start Building

When Does the Internship to Job Offer Pipeline Actually Work?

Here’s where I need to be honest. An internship is NOT a pipeline to a full-time offer. It’s more like an audition.

Most companies extend return offers to 30-40% of their interns. Some top-tier companies are higher (40-50%), some smaller ones are lower (10-20%). The point is, most interns don’t get offers.

This is not a failure. Let me repeat that. Not getting a return offer doesn’t mean your kid failed.

Why? Because return offers are for specific cases. When I looked at hiring decisions, I found three conditions that had to align.

The company has headcount - They need to hire and the intern is a solid engineer
The intern filled a gap the company knew it had - Maybe they’re hiring for Platform or Payments specifically and the intern impressed on that team
The timeline aligns - The intern finishes 2-3 months before graduation and the company hires for a start date in that window

Without all three? Even a great intern doesn’t get an offer.

I’ve seen brilliant interns at well-known companies not get offers because the company wasn’t hiring for their level when they graduated. I’ve seen average interns at smaller companies get offers because the startup desperately needed engineers and the intern was solid.

So when does the pipeline work?

If your kid interns at a company in the summer before senior year, does well, and that company is hiring for fall/winter new-grad positions, a return offer is likely. That’s the sweet spot.

If your kid interns earlier (freshman, sophomore summer), it’s less likely to lead to a return offer in the traditional sense. But it does lead to relationships with real engineers, a portfolio piece, and a reference from an actual employer. That’s still valuable. Real references and shipping proof matter more for their next opportunity than the offer itself.

How Do Internship Performance and Return Offers Connect?

This is where the rubber meets the road. What actually matters for a return offer?

From the hiring side, here’s what I looked at.

Technical competence (this is table stakes, not a differentiator). Can the kid code? Do they write reasonable code? Can they debug? If the answer is “not really” that disqualifies them. But if it’s “yes, solid junior-level” that’s enough.

Ownership - Did they see a project through? Or did they get stuck and need hand-holding? I wasn’t expecting independence. I was looking for someone who, when blocked, asked questions and moved forward. That’s huge. Most junior engineers struggle with this. Interns who have it are gold.

Communication - Did they keep their manager in the loop? Or did they disappear for a week and come back with something broken? This matters more than people think.

Attitude - Did they complain about the work being unglamorous? Or did they treat a debugging task like it mattered (because it does)? I’ve seen interns become jaded about “just fixing bugs.” The ones who own it, who ship it, who move on to the next thing. Those kids get offers.

The surprising non-factor is raw brilliance. I’ve seen incredibly talented interns who were prickly or didn’t communicate well. They didn’t get offers. I’ve seen solid, dependable interns who were great teammates. They got offers. The traits that matter in a 12-week internship are the same ones that matter in the first year of a real job. Shipping, clarity, reliability.

Here’s the thing parents should tell their kids. A return offer isn’t about being the smartest person in the room. It’s about being someone the team wants to work with next year.

Companies don’t hire students simply because they want to learn. They hire students because they have demonstrated the ability to build. — Why “Looking for an Internship” is the Worst Way to Find One

Does Brand Name Matter?

I can settle this. It depends but less than your kid thinks.

An internship at Google, Meta, or Apple absolutely helps. Here’s why. When your kid interviews at a different company the next year, that company assumes the candidate cleared a high bar. It’s a signal. A recruiter sees Google on the resume and thinks, “This person survived Google’s hiring process, so they can probably pass ours.”

But here’s the thing nobody talks about. That’s a signal in interview, not a signal that they’ll be a good engineer. After the first job, that signal fades fast.

I’ve hired people from Google and from “startups you’ve never heard of.” The Googler had an easier time getting interviews. But within three months of starting, nobody cared. What mattered was straightforward. Can they code? Do they ship? Are they easy to work with?

Now the flip side. Would my kid rather work at a startup than Big Tech? There’s actually an argument for it.

At a startup, an intern is more likely to ship meaningful work. They’re more likely to be trusted with actual problems. They’ll have fewer other engineers between them and the product. That’s better for your kid’s portfolio and your kid’s confidence.

At Big Tech, the intern might get more structured onboarding, better benefits, and a shinier company name. But they might be working on something internal that doesn’t ship or might be way down a chain of seniority.

Here’s my honest take. The best internship is the one where your kid actually ships something and learns something. The company name is secondary.

What Should Your Kid Actually Be Doing?

As a parent, here’s what you can actually influence:

Encourage multiple internships, not just one. Your kid should do at least 2-3 before graduating and working through school if possible. Each one is another signal, another portfolio piece, another reference. One internship is data. Two are a pattern. Three are a portfolio.

Once a student builds their first meaningful project, they’ve paid their technical debt. Every project after that becomes faster, more complex, and higher-signal—which is exactly what internships provide. — The 4-Year “Orchestrator” Portfolio Pipeline

Prioritize learning and shipping over the company name. A kid who spent a summer at a company nobody’s heard of but shipped a feature and learned how production systems work is better positioned than a kid at a famous company who shadowed people.

Ask what they shipped, not where they worked. When your kid comes home and says “I worked at Company X,” the question isn’t “Cool, is that prestigious?” It’s “What did you build? Can I see it?” If they can show you, you know the internship was real.

The gap between “I interned at Company X” and “I shipped a feature that reduced load time by 40% and here’s the GitHub PR” is where most internship applications die. — The Builder Delta: Why “Club President” Isn’t an Artifact

Encourage applying to smaller companies if they didn’t get Big Tech. The kid who gets rejected by Google but lands at a Series A startup often has a better internship experience. Less prestige, more impact.

Help them stay in touch with their internship managers. This is the underrated part. Years later, they’ll need references. They’ll need advice. They’ll want to know about job openings. The relationships they build in internships matter more than you’d think.

The One Thing to Remember

Here it is. An internship is practice for working. That’s it. It’s not a guarantee of anything. It’s not even a strong predictor of whether your kid will be a good software engineer. That’s usually about temperament and learning ability, not internship performance.

But it is the place where your kid learns if they actually like working as an engineer, learns how to work with real engineers, learns how production systems fail in messy ways, and learns how to ship. Those lessons are worth more than any company name.

Your child might have a perfect internship “credential,” but so does everyone applying. What actually moves the needle is shipping real work. — Confessions of a Tech Hiring Manager: Why I Ignore 4.0 GPAs

So when your kid is stressing about landing the “right” internship, remind them. There’s no perfect one. There’s just the one where they’ll work hard, ship something real, and come back having learned something. That’s the one that matters.

Have questions about internships or your kid’s career path? Drop them below. I read every comment and reply to as many as I can.

If your chid is a high schooler, The High School CS Internship Playbook I put together on Gumroad walks through high school level “how to get internship” process step by step, including the Builder Project framework, the portfolio review process, and real examples from students who were at this exact crossroads. You can find it at here.

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade. 👋

Support the Workshop: I spend my spare cycles auditing the “Source Code” of admissions so you don’t have to. If you’ve found value in these logic gates, you can buy me a coffee to help keep this workshop running and the guides open-source.

Support the Workshop

What AI Resume Builders Get Wrong About CS Internships

Tue, 14 Jul 2026 12:09:35 GMT

Parents ask me every week about AI resume builders and CV AI tools. Their student is applying for CS internships, the applications are disappearing into silence, and the logical next move looks like fixing the document. Run it through an AI resume builder. Generate a stronger CV. Polish the language. Try again.

These tools do exist and some of them are useful. But they solve layer two of a three-layer problem. Most students never get past layer one.

This is Part 3 of 6 in the College CS Toolkit preview series. Part 1 covered the identity shift, stop acting like a Student and start acting like an Engineer. Part 2 introduced the Bridge Projects, the two deployed builds that give you something real to show a recruiter. This part covers what to do with those projects in the application itself.

3.1 What AI Resume Builders Actually Do

An AI resume builder or CV AI tool does one thing well. It helps your resume pass an ATS, an Applicant Tracking System, by aligning your language to the keywords in a job description. If the posting says Python and your resume says scripting language experience, the resume builder flags that gap and helps you fix it. That is useful. It is layer two.

Layer two assumes you cleared layer one. But the reality is this, most students have not passed the layer one.

Layer one is framing. Before an AI resume builder touches the document, before the ATS scores it, a human somewhere makes a decision about whether your student’s application is worth routing to a real reviewer. That decision happens in two places: in the outreach message before the application, and in the three seconds a hiring manager spends on the first impression before scrolling.

An AI resume builder optimizes the document.

It does not fix the framing.

And the framing is what most rejections are actually about.

3.2 The Availability Trap

Ask a group of CS students what they are looking for and almost all of them give the same answer.

A software engineering internship in this area.

The city changes. The sentence structure stays the same.

That sentence describes what the student wants. It says nothing about what the student can offer the team receiving the message.

Hiring managers respond to problems they can solve.

They do not respond to availability they need to accommodate.

The student who writes I am looking for a software engineering internship is handing the company the work of figuring out where the student fits.

The student who writes I built a live monitoring dashboard that reads sensor data and deploys to the cloud and I want to bring that to how your infrastructure team handles X is handing the hiring manager a reason.

Same background, same resume, same AI resume builder output.

The framing is the only variable. Different result.

I call this The Availability Trap. Most students fall into it because nobody named it for them. The framing runs before the robot, before the recruiter, before anyone reads a single line on the CV.

As I laid out in Why Looking for an Internship is the Worst Way to Find One, companies do not hire students because they want to learn. They hire students because they have demonstrated the ability to build.

The shift from Consumer framing to Builder framing is the foundation this entire application strategy builds on. As I covered in The Builder vs. Consumer Framework: Why Your Perfect Resume Is Not Enough, students with perfect stats get waitlisted or rejected not because the document is wrong but because the posture it signals is wrong. A better CV AI output does not fix posture.

3.3 The Robot Layer

Once the framing is right, the AI resume builder becomes genuinely useful. The ATS scores your resume against a keyword list pulled from the job description.

If the score is below the threshold, no human ever opens the file.

A resume builder that aligns your language to the posting raises that score.

This is mechanical, not unfair. The ATS does not read engineering judgment. It matches strings. A resume that says built a REST API in Flask with live deployment scores differently than a resume that says worked on backend systems.

Same student, same project, different language, different score. A good AI resume builder or CV AI tool catches that gap.

But there is a second way past the ATS that no resume builder can produce for you.

A warm referral from inside the company routes your resume directly to a hiring manager.

The ATS queue is bypassed entirely. Twenty plus years of hiring across Intel, HP, Toyota, and AI programs taught me this consistently. Warm referrals get read. Portal submissions get scored. Those are not the same outcome.

As I described in Confessions of a Tech Hiring Manager: Why I Ignore 4.0 GPAs, I once had 150 resumes on my desk from target schools. Almost all listed club president roles. The ones I moved forward were the ones where someone I trusted said this candidate builds things. That is the warm referral in practice, and it happens before the AI resume builder ever enters the picture.

3.4 The Warm Outreach Formula

The average CS applicant avoids outreach because they do not know what to say. The message they imagine sounds like begging, so they run the resume through a CV AI tool instead and submit through the portal. The portal has a drop rate most students never see because it happens before any human reviews the file.

You do not need to know anyone at the company before you reach out. You need to know one specific thing about what the team is working on.

Three sentences.

Start with one specific observation about the team, the product, or the person’s work.

Not a compliment, a fact that proves you looked.

I read about the caching overhaul your team shipped last quarter is more useful than I really admire your company’s engineering culture.

Follow with one sentence about what you built. Not the AI resume builder summary, one project, one outcome, one link. I built a Flask API with live sensor integration deployed to Render and the GitHub repo is public.

Close with one specific ask. Not can I have an internship, but can I have a fifteen minute conversation with someone on your team about what the summer internship looks like.

Specific observation, specific artifact, specific ask. Students who write eight sentences and say less are the ones who never hear back.

3.5 What the Projects Make Possible

The Bridge Projects from Part 2 are the artifact in sentence two of that message. A live deployed project with a real URL gives you something concrete to name. That is why those projects exist. Not to impress a judge, not to make the AI resume builder summary sound better. To give you something specific to say to a real person at a real company.

As I covered in The Builder Delta: Why Club President Is Not an Artifact, the gap between listing something and building something is exactly where most applications die. The warm outreach formula requires a real artifact. A CV AI tool cannot generate that artifact for you.

Students who do not know the sequence run their resume through a builder, submit through the portal, and wonder why nothing comes back. The ones getting callbacks are sending three sentences to real people and naming one real thing they built.

For the full action plan on converting the Builder identity into an active application strategy, see Pre-Summer Action Plan: From Consumer Trap to Builder in 8 Weeks.

Part 4 covers what happens after you get a response.

If you want the full six-part framework in one place, the College CS Toolkit covers everything from the identity shift to the offer decision. Start with this free preview series and see if the approach makes sense for your student first.

Masaru
The Engineering Dad

Have questions about internships or your kid’s career path? Drop them below. I read every comment and reply to as many as I can.

If your chid is a high schooler, The High School CS Internship Playbook I put together on Gumroad walks through high school level “how to get internship” process step by step, including the Builder Project framework, the portfolio review process, and real examples from students who were at this exact crossroads. You can find it at here.

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade. 👋

Support the Workshop

A Fake Account Just Posted My Advice Using an AI Rewriter. I Caught It Because the AI Made One Mistake.

Thu, 09 Jul 2026 11:48:02 GMT

Something showed up in one of the college admissions Facebook groups last week that I want to tell you about.

A comment appeared under a parent question about whether a rising CS senior should double major.

The advice was solid.

It mentioned building a real minimum viable product to GitHub before applications open. It referenced using modern AI APIs from OpenAI and Anthropic.

It sorted school choices into High Reach, Reach, Target, and Safety, while most college counselors use Reach, Target and Safety.

It opened with a phrase about being a father of three computer science graduates speaking from a tech background.

I stopped.

I had posted almost that exact advice in the same thread the day before.

The account name was Jenny Lozada. Female name. Profile photo of a woman. The bio said she follows Financial Education for good tips. But the comment opened with Father of Three Computer Science Graduates Here.

The AI rewrote every sentence.

It forgot to change that one word.

I reported it out to the FB group admins, and it has since been taken down.

How AI-Generated Content Theft Works in Admissions Groups

This is not a brand new trick, but it is spreading fast and getting harder to spot. Someone finds a comment that resonates, a real answer from a real person who actually knows what they are talking about, and they run it through an AI rewriting tool. The tool keeps the ideas, restructures the sentences enough to avoid a simple copy detection search, and the output goes up under a fake or compromised account in another group or the same one.

The Jenny Lozada account had never posted in that group before. Her comment was her debut. The bio pointing toward Financial Education is an engagement farming signal, not a parent looking for admissions help. Real contributors in these groups leave a trail. Real parents have history, follow-up questions, kids whose grade levels you can track across threads. Jenny Lozada had none of that.

I witnessed a fascinating and dangerous interaction in a major parent group this week. A commenter was confidently telling a group of 50,000 parents that Northeastern caps its merit aid at $5,000.
— The $20,000 Rumor: Why Facebook Experts Are Costing You Money

What AI-Generated Advice Cannot Copy

What caught me was not just the gender slip. It was the intellectual fingerprint.

The sequence of ideas in that comment is mine.

The anti-double-major-in-unrelated-fields stance.

Hardware alongside software.

Build a functioning MVP with real AI APIs before applications open.

Tier the school list into High Reach, Reach, Target, Safety.

That specific combination comes from 15 years of doing tech hiring, from coaching robotics teams from FLL to FTC to FRC and World Championships, from watching three kids choose Cornell, Rice over Caltech and Columbia on full ride scholarships.

It is not a sequence you arrive at generically.

It is built through repetition.

AI can rewrite a sentence.

It cannot rewrite the 15 years that produced the sentence.

And that is exactly the point. The stolen version gave the surface answer. But it has no ability to handle the follow-up.

If you spend any time in online career groups, you will see hundreds of variations of the exact same post. It rarely leads to legitimate opportunities and almost exclusively attracts bots, spammers, and pay-to-play programs.
— Why Looking for an Internship Is the Worst Way to Find One What if the student has zero hardware background. What does a GitHub MVP that impresses a real hiring manager look like versus one that looks like a class assignment. Which AI APIs are worth building on now versus which ones are getting abstracted away next quarter. The real thinking can go there. The AI-generated copy cannot.

What This Means for Parents Reading Admissions Groups

These groups are genuinely valuable. Real experts are in them, people with actual track records and actual kids. The noise is getting louder but the signal is still there.

Here is the filter I would use.

When AI-generated-sounding advice appears from an account with no group history, ask one follow-up question.

Something specific to your student’s situation.

Real experts go deeper.

They get granular, push back on edge cases, offer the version of the answer that only applies to your kid.

AI-laundered advice tends to go wide when pressed. It gives you more of the same surface answer because there is no actual judgment behind it.

Also watch the bio. Legitimate parents in these groups are not promoting Financial Education channels.

That is a spam signal wearing a human costume.

What This Means If Your Own Content Gets Copied

When someone takes your content and launders it through AI, that is uncomfortable to discover.

It is also a signal worth reading.

It means you said something specific enough to be worth stealing.

Generic advice does not get stolen because there is nothing distinctive to take. If someone found my advice worth manufacturing a persona around, I said something that cut through the noise.

I spent about twenty minutes crafting a detailed, data-driven response to a parent whose student was feeling undecided. I shared the framework used to find students’ North Star and mentioned the guide I put together to help others do the same. But then I realized that this skepticism is actually a symptom of the very problem I’m trying to solve.
— Why I Give This Away (And Why Some People Hate It)

The protection is not watermarking or locking down your posts.

The protection is what I write about in The Builder Delta.

The artifact layer.

A published body of work, a consistent presence over years, a track record that connects every piece of advice to something real and verifiable.

Jenny Lozada posted one comment and disappeared.

I have been in these groups for years.

That asymmetry is not copyable.

Because they are world-class Consumers, but the Ivy League is looking for Builders.
— The Builder vs. Consumer Framework: Why Your Perfect Resume Isn’t Enough

Build the artifact layer. Keep showing up. The copies fade. The original compounds.

The gap between those two sentences is where most applications die.
— The Builder Delta: Why Club President Isn’t an Artifact (And How to Fix It)

If you want the complete framework on building a portfolio that cannot be faked, the High School CS tool kit (and the sequel of College CS Internship Playbook is being released) on Gumroad walks through exactly this.

You can find it here.

And if you have seen something similar in the groups you follow, reply to this post. I want to understand how widespread this is becoming.

No pitch, no spam. Just the same conversation I have been having in parent groups for the past decade. 👋

Support the Workshop

Your To-Do List App Is a Rejection. Build These Two CS Projects Instead.

Tue, 07 Jul 2026 12:00:01 GMT

Last Tuesday I opened this series with the identity shift. Stop acting like a Student and start acting like an Engineer. If you missed Part 1, start there.

This is Part 2. The Bridge Projects. The two deployed builds that turn last week’s identity shift into proof a recruiter can click. If you have ever searched for CS projects for a resume and ended up with a list of fifty tutorial ideas, this part exists to replace that list with two.

First, the uncomfortable diagnosis. Most students are stuck in what I call Tutorial Hell. Watch a video, copy the code, end up with a To-Do List app running on localhost. Then it goes on the resume next to the GPA, and the student waits for interviews that never come.

The app is not the problem.

The pattern it signals is.

A tutorial app tells a hiring manager you can follow instructions. Last week I explained why following instructions is the one skill the market already has too much of.

2.1 The Golden Rule of Resume Projects

If it is not live, it does not exist.

The Amateur says it works on my machine.

The Pro says it runs in the cloud.

Every project you list must clear two bars.

A live URL. Hosted on the web, like my-app.onrender.com. A recruiter who clicks it sees it running.
A public GitHub repository. The code, the commit history, and the documentation, all visible.

That is the entire rule. It filters out every tutorial app, every class assignment sitting in a zip file, and every project that dies when the laptop closes. In Confessions of a Tech Hiring Manager I walked through the resume stacks I reviewed for fifteen years. The projects I could open in a browser got my attention. The ones I had to take on faith got skipped.

“In today’s highly competitive market, companies do not hire students simply because they want to learn. They hire students because they have demonstrated the ability to build.”
— Why Looking for an Internship is the Worst Way to Find One

2.2 Before You Build. The SOS Protocol

Here is the objection I hear from parents before I even finish describing the projects.

“AI writes all the code now, so a project proves nothing!”

Half right. AI can produce the code. That is exactly why the toolkit opens Part 2 with the SOS Protocol before a single line gets written. The protocol is a set of Universal Tutor prompts that turn the AI from a vending machine into a tutor, and the difference decides whether the project survives a technical interview.

Three pieces preview well.

The Explain It Back Ritual. You cannot paste AI code into your editor until you have explained it back, line by line, and been corrected. If you cannot explain a line, you delete it.

The Debug Loop Rule. When you hit an error, the AI must quiz you on why it happened before it shows you any fix. Pasting the error and copying the solution teaches you nothing, and the interview will find that out.

The Offline Challenge. Before you ever interview, explain your project to a rubber duck, or your dad, without looking at the code. If you cannot draw the architecture on a napkin from memory, the AI built it. Not you.

As I covered in The $0 Vibe Coding Stack, AI-assisted building is the on ramp now, and the students who win are the ones who use it to learn faster instead of to skip learning. The full guide has the complete prompt scripts, including the constraint lines that force the AI to refuse when you try to shortcut it.

2.3 Project A. The Job Tracker API

The first of the two CS projects for your resume is a backend build. Roughly 10 focused hours for a student comfortable with basic syntax. The goal is to prove you understand data persistence, scalability, and system design.

Why an API and not a website.

The API is the nervous system of the internet. Most modern software never touches a human. Systems talk to systems, the way Uber talks to Google Maps. And when an interviewer asks you to design a URL shortener, they are asking you to design an API. Building one now means rehearsing the exact skill the whiteboard will test.

The mindset shift matters more than the stack.

Do not build a tool.

Build a SaaS.

Most students build apps that work for one user, themselves.

Design this one as if you will sell it to every student on campus.

A relational database with indexing, architected for thousands of users.

One warning from the guide, and it is not optional.

Never claim you have those users.

Say you designed for that scale.

The first sentence is a lie. The second is Business Acumen.

The stack is deliberately boring and deliberately industry standard. Spring Boot or FastAPI, PostgreSQL, deployed on Render or Railway. Boring is the point. These are the names that appear in the job descriptions.

The guide walks the whole build through AI protocols modeled on how a real team works, including a Roast and Refactor method where you write the ugly first draft and the AI critiques it like a lead developer. It ends with a final exam called the Backend Orchestrator, where the AI plays a skeptical senior lead and fails you for the two deaths recruiters actually see.

The Cold Start Death, a live URL that spins for more than ten seconds before loading.

And the Persistence Fail, a server restart that wipes your data because the database was never wired up correctly.

2.4 Project B. The Campus Market

Project A is invisible. Project B must be beautiful.

The second build is a marketplace where students buy and sell textbooks and furniture. Craigslist for your campus. Roughly 14 hours baseline, built on Next.js, Tailwind, and Vercel, the de facto production stack for modern frontend work.

The real lesson hiding inside Project B is not the interface. It is the workflow. In a real company you never push code straight to the main branch. That is how production breaks. The guide makes you work the Feature Branch Strategy from the first commit. Main is production and must always work. Your feature branch is the sandbox where you make your mess. The Pull Request is how you ask permission to merge. A junior who does not break production is rare, and the commit history proves the habit before anyone asks.

My favorite piece of Project B is the Twist. The guide has the AI ask about your non-coding interests, then derive one Functional Business Feature from the answer. Logic, not decoration. A gamer does not make the background neon. A gamer adds an XP system for every item listed. That single feature is what separates your Campus Market from the other four in the recruiter’s tab history, and it gives you a story in the interview that no tutorial can produce.

Project B closes with its own exam, the Deployment Orchestrator, plus a Day Zero Maintenance Checklist. The incognito test, the mobile test, and the Database Awake test, because free hosting tiers put databases to sleep and a recruiter will not wait for yours to wake up.

If a student wants to strategically deviate from Project B, there are three specialization paths the student can pursue. Get the full version to see details.

2.5 The 24 Hour Objection

Two builds, roughly 24 hours combined.

Here is the pushback.

“My student is buried in coursework, there is no room for 24 hours!”

The honest math is worse than that, and the guide says so up front.

The estimates assume a “Seasoned Student”.

If your student is learning the language while building, the multiplier is 2x to 3x.

Project A might take 30 hours instead of 10.

The guide calls that tuition, and I stand behind the word.

That is not wasted time.

That is the necessary cost of bridging the gap between the University Stack and the Industry Stack, and it is the cheapest tuition your family will pay this decade.

Spread across a semester, it is two focused weekends a month. If your student is still in high school, The Pre-Summer Action Plan sequences the same shift across 8 weeks before freshman orientation, and as I laid out in The Builder vs Consumer Framework, the students who make it are the ones who moved from consuming credentials to building artifacts before the deadline forced the question.

What Comes Next

This preview showed you the two builds.

The full guide gives you the execution layer.

The prompt scripts, the schema decisions, the deployment walkthroughs, and the exams that verify the work.

Next Tuesday is Part 3. The Application Strategy. Your resume will be read by a robot before any human sees it, and most CS resumes lose to the robot. Part 3 covers the ATS gauntlet and the Ugly Resume Paradox.

Drop a comment with one answer. If your student started this weekend, which build comes first, the invisible Job Tracker API or the beautiful Campus Market. Your reasoning tells me more about where they are than any GPA could.

Let’s build,

Masaru
The Engineering Dad

If your chid is a high schooler, The High School CS Internship Playbook I put together on Gumroad walks through high school level “how to get internship” process step by step, including the Builder Project framework, the portfolio review process, and real examples from students who were at this exact crossroads. You can find it at here.

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade. 👋

Support the Workshop

CS Major Can't Find a Job. Neither Can the Chemistry Grad. The Major Was Never the Variable.

Thu, 02 Jul 2026 11:23:03 GMT

Two posts landed in my feed this week from two different parent groups.

Post one. A mom in the Parents of Class of 2026 group. Her son just graduated with a bachelor’s degree in chemistry and physics, minor in mathematics. He cannot find a job. She asked the group for suggestions. He had looked at medical lab positions but did not qualify. No biology on his transcript.

Post two. A parent in a College Admissions group. Her son wants to pursue AI product development. Rising senior. 4.0 unweighted, 4.4 weighted. 1470 SAT with an 800 in math. Research with a university professor this summer. Coding tutor for 15 months. Varsity athlete. She is wondering whether CS is oversaturated and whether to double major for options.

I read both posts carefully.

Neither one mentioned what these students built.

Not one project. Not one GitHub link. Not one tool created for a real user. Not one problem solved for someone outside the classroom. Just degrees, grades, activities, and the worry that the major might be the wrong one.

Here is what I have seen coaching robotics teams from FLL, FTC to World Championships, and watching three of my own kids choose Cornell, Rice, over Caltech and Columbia on full scholarships, and NEU.

The major is almost never the variable.

The artifact is.

Why a CS Major Can’t Find a Job. And Why a Chemistry Major Can’t Either.

The chemistry and physics graduate is not stuck because of his major. He is stuck because his job applications look identical to every other chemistry and physics graduate. Degree. Coursework. Grades. No evidence of what he can actually do when a real problem is in front of him.

The rising CS student is not going to solve his problem by adding a second major. If he ships nothing in CS, he will ship nothing in CS plus mathematics or CS plus business. The extra credential does not change the signal he sends to a hiring manager.

I spent over 15 years doing tech hiring. 150 resumes on my desk at a time. 130 with a 3.8 GPA or higher. Most from target schools.

I deleted the majority of them.

What I was looking for was the student who could point to something real that I can see and test with. Not I have experience with Python. Something like I built this tool for our lab. Forty students used it. Here is the GitHub. That gap is the entire ballgame.

I had a physical stack of about 150 resumes sitting on my desk. 100 of them were from Target Schools. Almost all of them listed President of Club or Member of Society.
— Confessions of a Tech Hiring Manager: Why I Ignore 4.0 GPAs

The Builder Delta

I wrote about the three signal levels in The Builder Delta. Short version here.

Level 1 is awareness. You know a problem exists.

Level 2 is a personal solution. You built something, but only you use it.

Level 3 is adoption. Other people use what you made without you pushing them.

The gap between those two sentences is where most applications die.
— The Builder Delta: Why Club President Isn’t an Artifact (And How to Fix It)

Most job seekers, CS majors included, are sitting at Level 1 or 2. The market right now is brutally efficient at filtering this group out. The students who are moving are at Level 3. And you do not reach Level 3 by picking the right major. You get there by shipping real things to real people and documenting that they actually used it.

This is true whether your kid studied CS, chemistry, biology, or English.

A high-achieving student presents a resume that is virtually perfect: 150+ volunteer hours, President of three clubs, Varsity athletics, and a list of summer programs that would make a Fortune 500 CEO sweat. And yet, many of these students are met with Waitlisted or Denied from Tier-1 institutions like MIT, Cornell, or Stanford.
— The Builder vs. Consumer Framework: Why Your Perfect Resume Isn’t Enough

What the Chemistry Grad Could Have Done

Chemistry and physics touch everything. Environmental sensors. Lab automation. Drug formulation workflows. Materials testing software. If this graduate had spent two summers building a working tool for any of those domains, even a free small one, and if real users had touched it, he would not be asking the group for suggestions right now. He would be pointing to something.

The degree was not the problem. The empty portfolio was.

What the CS Applicant Still Has Time to Do

The rising CS student has the ingredients. Professor research. 15 months of tutoring experience. Strong academics. But the post describes him as someone with an interest in AI product development. Not an AI product.

If he ships something real this summer, something a teacher, a coach, or a peer can actually use, CS oversaturation stops mattering by application season. The artifact answers the major question before anyone even asks it. I wrote more about why this matters in The Post-Pure-CS Era.

In today’s highly competitive market, companies do not hire students simply because they want to learn. They hire students because they have demonstrated the ability to build.
— Why Looking for an Internship Is the Worst Way to Find One

The Builder Project

This is what I tell every family I work with.

Do not build a passion project. Build a Builder Project.

A passion project starts from what your kid loves. A Builder Project starts from a real problem someone else has. You find the friction. You build something that removes it. You document that other people adopted the solution.

That sequence is the credential. The major is the language you used to get there. Any major can produce a Level 3 builder. Most majors produce consumers who cannot find a job regardless of the field.

Back in my college days, I double-majored in CS and Business because I knew a Business degree by itself wouldn’t be enough to survive in the tech industry. In 2026, that reality is no longer just for the tech industry.
— The End of the Generic Business Degree (And How to AI-Proof Your Finance Kid)

The High School CS Internship Playbook I put together on Gumroad walks through this entire process step by step, including the Builder Project framework, the portfolio review process, and real examples from students who were at this exact crossroads. You can find it at here.

No spam, no pitch. Just the same conversation I have been having in parent groups for the past decade. 👋

Support the Workshop