About
I’m a software builder focused on AI-assisted development and practical agent systems. I care about shipping reliable features, improving developer experience, and turning research into workflows teams can use today.
What I do
- Design agentic workflows for real apps (retrieval, tool use, evaluation)
- Bridge product and engineering to move from prototype to production
- Share practical patterns for GitHub Copilot, RAG, LangChain, and data choices
- Advocate human-in-the-loop systems that are observable, testable, and safe
Principles
- Start simple, measure, iterate
- Humans stay in the loop
- Shipping > demos
- Privacy and security by default
Currently exploring
- Agents for production developer workflows and CI
- Copilot-native team practices and code review
- Evaluation harnesses, datasets, and telemetry for AI quality
A bit more
- I write playbooks to operationalize LLMs: prompts → tools → evals → shipping.
- I prefer boring, composable systems that scale with observability and tests.
- I care about clear docs and small feedback loops more than shiny demos.
Connect