Software engineer with fifteen years of experience, mostly in Python and Django on Google Cloud. Lately I've focused on AI engineering: agent harnesses, hooks, and workflows with Microsoft Copilot and Claude Code. Previously I was a lead instructor at Galvanize for two years, leading and teaching eight cohorts of full-stack JavaScript. I still use React and TypeScript on side projects.

For the past five years I've been the lead developer at Koinonia Family Services, maintaining Django applications used by social workers to care for foster families. They support thousands of placements and roughly 200,000 days of care a year. They run on Kubernetes and App Engine, and I manage EHR integrations, HIPAA compliance, and the GitLab CI/CD pipeline I built to test and deploy them.

I'm looking for a new role on a remote team. I'm excited to learn different stacks, and I love partnering with stakeholders, building products, and mentoring developers. Happy to be onsite for gatherings, while staying rooted here in northern California. I'm open to any awesome humans doing good things. Please feel free to reach out!

• In production now: Python, Django, JavaScript, PostgreSQL, Docker, Kubernetes, Google Cloud (GKE, App Engine, Tasks), GitLab CI/CD, LLM agent harnesses (Claude Code, Copilot)
• In production earlier: React, Vue, Angular, Node.js, Go, Flask, Jupyter, pandas
• On side projects: FastAPI, SQLAlchemy, Strawberry GraphQL, React, TypeScript, RAG (embeddings, hybrid search, reranking, evals)

I had a contribution merged into Django that had been requested since 2010.

Most of my work is closed source, but here are some public repos:

• Secondhand Plastic (source) matches secondhand food containers with people who'll use them, instead of the bin. Server-rendered FastAPI and Jinja2 on Supabase Postgres, no frontend build step.
• The Mullet Stack (source) is a field guide to wiring modern Python and JavaScript together, built around one feature shipped across both. The frontend's types are generated from the backend's OpenAPI schema, so the two ends can't drift. FastAPI, Pydantic, React, TypeScript, Vitest.
• RAG in context (source) is a notebook refreshing my own understanding of retrieval-augmented generation and connecting it to the older ideas it rests on: searching by meaning and by keyword, why you want both, rank fusion, a reranker, and 18 labelled questions to check the whole thing against. No API keys, runs offline on a laptop.
• playground is a habit tracker in Go and SQLite, installable as a PWA, deploying itself to Fly.io.

Lately my time has gone into writing and Cohort 10 of the Climatebase Fellowship.

I like helping people feel welcome in communities. I co-chaired the sprints at DjangoCon US 2025, captained teams in Djangonaut Space sessions 5 and 6, founded Central Coast Python, and co-organized BoulderJS before that.

I'm a surfer, meditator, and permaculture practitioner. I earned a certification in abundant food forests at OAEC and completed a 200-hour yoga teacher training.

[email protected] · seanhelvey.com · LinkedIn · Bluesky · Mastodon · Goodreads