Series

Building AI Products: Engineering Notes from the Field

Real problems, experiments, trade-offs, and lessons learned while building AI-powered products.

Oct 19, 2026
Oct 19, 20268 min read2 views
A Passing Eval Doesn't Mean Your AI Feature Is Reliable
How I think about evaluation sets, failure analysis, and the gap between a good score and a dependable AI product.
8 min read2 views
Oct 15, 2026
Oct 15, 202614 min read13 views
My LLM Judge Went From 70% to 96% Agreement With Humans. Here’s What I Learned.
What three iterations of an LLM-as-a-judge rubric taught me about evaluation design, regression testing, and synthetic data quality.
14 min read13 views

Building AI Products: Engineering Notes from the Field