Field notes from running AI programs in large enterprises. Honest, specific, occasionally opinionated.
3 tiers, 24 measures, a hard safety gate, and a triage system for figuring out not just that something failed — but why, and what to fix.
Read more →Exploring a RAG + few-shot prompting pattern for test case generation — with a feedback loop that gets smarter over time. Partly built, partly hypothetical, fully worth thinking about.
Read more →