Andrew Brook

I build software and write about what the data actually said. Lately: deterministic libraries for the things LLMs get wrong, benchmarks that test agents the way they're really used, and voice-first applications.

GitHub · LinkedIn · RSS

Writing

Nothing published yet.