I'm an end-to-end AI builder focused on evaluations, benchmarks, and verification β making AI systems more rigorous and trustworthy. Currently Principal AI Engineer at Curebase, bringing evaluation to AI in clinical trials.
My background spans neuroscience research (Yale, SRI International), data science, and software engineering. I believe evals are the weak link in AI development, and I'm working to change that.
- π€ Speaking at meetups and conferences on AI evaluation practices ("The Eval Flywheel", "Evals, Benchmarks, and Guardrails")
- π¨πΏπ¬π§ Co-founder of the Evals.cz meetup, running in Prague and coming to London
- ποΈ Co-hosting the Data Talk and AI ta Krajta podcasts
- π©βπ« Teaching Data & AI courses at Czechitas
- Don't Write Evals for Fast-Moving Systems (2026)
- Clobsidian in Detail: Cross-Source Personal Infrastructure (2026)
- Clobsidian, and other winter experiments with Claude Code (2026)
See my blog for a full list of articles.
- Cognitive Exhaust Fumes, or: Read-Only AI Is Underrated β AI Engineer Europe (online track), Apr 2026
- Evals, Benchmarks, and Guardrails: A Pythonista's Guide to Not Mixing Them Up β PyData Prague, Feb 2026
- When (& How) to Start Writing Evals β Evals.cz Meetup #1, Feb 2026
- Simultaneous Vibe Coding β TopMonks CaffΓ©, Jan 2026
- Pydantic, Everywhere, All at Once β EuroPython Prague, Jul 2025




