Now

Open source

Contributing to AI/agent infrastructure repos — working through a prioritized list of real bugs in projects I actually use: mem0, pydantic-ai, instructor, vercel/ai, and the Google ADK JS SDK. The goal is merged PRs on things that matter, not doc typos.

My approach: find unassigned bugs with reproduction steps, confirm locally, fix with a test, open PR. Targeting repos where maintainers review outside contributions within a week — low crowd, fast feedback loop.

Job search

Looking for a remote role — full-time or contract. I work across the stack: AI integrations, RAG pipelines, agent systems, backend infra, fullstack product. Open to any team building something real where I can own meaningful parts of it.

If something I've built is relevant to what you're working on, reach out.

Reading

Working through codebases more than papers right now. Understanding how production agent frameworks handle state, tool calls, and failure recovery at the implementation level — not just the API surface.