AI product manager. I ship agentic and LLM products from zero to one, and I own the evals and observability that make them trustworthy. The tools below are how I work.
Sole PM at SignalQuest for GPS products and a companion app used by earthworks contractors. Dartmouth MEM. Electrical engineer by training.
- llm-peer-review: multi-model AI debate for your entire project lifecycle, plans, specs, research, and code, for product managers and small teams who build with AI and want the plan questioned before the code exists. Every repo here was built with it.
- gitmargin: Google-Docs-style comments on HTML prototypes, returned as a batch a coding agent can apply in one pass. No server needed. Live demo.
- trace-annotator: keyboard-first open coding and error analysis on LLM traces, built for PMs starting evals. Try it at trace-annotator.vercel.app.
- pokedex: chat-based product lifecycle management where the AI proposes and you confirm before anything is written. Live at pokedex-plm.vercel.app.
LinkedIn. For anything about these repos, open an issue.