- LLMKit - AI gateway and SDKs for measuring agent cost, enforcing budgets, and controlling sessions.
- f3dx - Rust/Python runtime prototype exploring agent execution, telemetry, replay, and caching.
- tracewright - Evaluation prototype for replaying traces and comparing behavior across test cases.
- Pydantic Monty - upstream contributor with four merged fixes across runtime safety, parser correctness, error propagation, and diagnostics. View merged work.
Open to AI systems, SDK, platform reliability, and evaluation engineering roles.




