Aarush Sah

I work on AI. For the past few years that has mostly meant evaluation - deciding what to measure, and building the infrastructure to measure it properly.

When I was at Groq, I built openbench - an open-source framework for running language-model benchmarks easily and reproducibly.