Work

AI research projects.

Agentic evaluation and observability for production LLM systems running across every frontier model — OpenAI, Anthropic, Gemini, and self-hosted. Worked across Python/Quart backend, React/TS frontend, and three client SDKs (Python, TypeScript, JVM). Built tools for surfacing insights from production agent runs and the iteration loop that helps teams actually improve those agents.

Test-suites APIs and dashboard for a time-travel debugging platform. TypeScript, Node, Postgres, GraphQL, Kubernetes. Zero-downtime migrations and observability tooling. Promoted from intern.

Open-source full-stack JS framework. Build tooling, CLI generators, test infrastructure. Cut Jest heap memory by roughly 1.7×.

Built and shipped product work for several teams — notably Tape.sh, Planwise, and teamstream.gg. RedwoodJS, Next.js, Electron, ffmpeg pipelines.

Launched the offers and promotions microservice for an Indonesian digital bank. High-availability BigQuery streaming for batch notifications.

Location-aware audio guide for tourist attractions. Led a team of six. Invited to Startup India Global VC Summit 2019.