Building & teaching
with AI agents.
Applied AI Architect @ Anthropic
I design agentic systems — and help people learn to build with them. Mostly I’m here to make the hard parts make sense.
I work at the seam between people and capable software. As an Applied AI Architect at Anthropic, I help teams turn frontier models into agents that actually do the work — reliable, well-scoped, and trusted enough to hand real tasks to.
Almost everything I know, I learned by building it myself — usually solo, usually from scratch. Home systems that run themselves, pipelines that research and write without me in the loop, small agents that quietly handle the busywork of a day. I build in the open — sharing the wins and the dead ends as I go — and I teach the same way, on camera and in writing, because the fastest way to understand a thing is to build it, then explain it to someone else.
I’m less interested in demos than in the unglamorous middle: the guardrails, the feedback loops, the moment an agent stops being a toy and becomes something you’d trust with your morning.
- Agents are the new interface. The next decade of software won’t wait to be clicked — it will act, check its work, and report back.
- Seek truth; do right when no one’s looking. Integrity isn’t situational. Follow the evidence, own the mistakes, and act the same whether or not anyone is watching.
- Clarity beats cleverness. A system you can’t explain is a system you can’t trust. Make the hard parts make sense.
- Build in public, learn in public. The best feedback loop is shipping something real and watching what breaks.
- Leverage should be shared. The point of automating the busywork is to give people back their attention for the work that matters.