Open-source contributor to OpenAI and Microsoft projects. I build developer tools and dependable AI systems, with an emphasis on reproducible fixes, security boundaries, and software that is useful outside a demo.
- AI agent infrastructure — orchestration, memory assurance, remote agent interfaces, and evaluation.
- Developer tooling — CLIs, browser-first utilities, code-review workflows, and automation.
- Reliability and security — regression tests, cross-platform behavior, filesystem safety, privacy boundaries, and explicit failure modes.


