AI That Ships

The gap between building AI and running it. Production notes on ML, data science, and the systems in between — handoffs, evals, cost, and what actually survives contact with users.