Sebastian Fox is the CEO and co-founder of Composo, building an LLM evaluation layer that uses generative reward models to achieve state-of-the-art agreement with human expert judgment (~92% accuracy with very low variance) — well above typical LLM-as-judge approaches. Attending his session offers a grounded, production-tested perspective on why standard evals fail in clinical and agentic AI systems, and what it actually takes to catch failures before they reach end users.

Public activity researched automatically · as of Jun 2026