Software Engineer
About
I build systems that keep working after the demo — agents, APIs, and the careful parts in between.
What I do
I'm a software engineer. Most of what I build sits between a model and a production system: agents that call tools, APIs that have to answer in milliseconds, and the queues, budgets, and kill switches that keep both honest.
I like projects where the boring parts decide the outcome. Nash AI reviews pull requests on GitHub, Gentle Stream ranks a news feed without calling a model on the request path, and Stateful Indecision keeps an agent's every decision in a hash-chained ledger. The case studies cover the tradeoffs.
Skills
- Languages
- Python, TypeScript, SQL
- Frameworks and libraries
- FastAPI, Next.js, React, SQLAlchemy, Pydantic, Astro, Tailwind CSS
- Data
- Postgres, Redis, Supabase, Alembic migrations
- AI and agents
- Anthropic, OpenAI, and Gemini APIs, Tool-using agent loops, Eval harnesses, MCP servers, Langfuse tracing
- Infra and delivery
- Docker, GitHub Actions, Cloudflare, Vercel, Railway, Sentry
- Quality
- pytest, Vitest, Playwright, Ruff and mypy, Bandit