About

I spend most of my time right now figuring out how to make AI coding agents actually work in enterprise environments, not toy demos, but reliable systems that teams can trust with real codebases. The gap between "impressive demo" and "production-ready tool" is enormous, and that's the space I'm obsessed with.

My current focus is on spec-driven development for agents: giving them structured specifications instead of vague prompts, and building verification frameworks that catch mistakes before they hit production. Building a systematic approach to validating agent output that goes way beyond "does it compile." Think type checking, architectural conformance, security scanning, test generation, and more, all orchestrated as guardrails around agent workflows.

I also work on the infrastructure side: how agents can autonomously provision and deploy to cloud platforms, and how we build web applications that work seamlessly for both humans and AI agents. I've been sharing these ideas at conferences like DotAI, SymfonyCon, and AI Day, and through live coding sessions where I put agents to the test on stage. It's a wild time to be building this stuff, and I genuinely believe we're only scratching the surface.