Now
What I'm building
AI-enablement work with software teams. Agentic harnesses, test infrastructure, self-verifying loops. And my own tooling: the harness I benchmark in public on a 16-year-old app.
What I'm writing about
Agentic engineering: what it takes for AI to ship production code. Harnesses, self-verification, and why the model was never the missing piece.
What I'm exploring
Homelab infrastructure. Self-hosted services, monitoring, cross-site networking.
What I'm open to
A small number of AI-enablement engagements. Hands-on, with your team, in your codebase. Currently open. And conversations with anyone seriously building with AI agents.