How I build

I'm an independent engineer who builds AI systems end to end: the product idea, the architecture, the rules the AI must follow, the tests that decide when something is done, and the servers it runs on.

I don't hand-write every line. Coding agents do a large share of the implementation and review. They work inside a harness I designed: a shared knowledge base, defined roles, stage gates, independent reviewers and a human approval before anything ships. The products, the designs, the acceptance criteria and the final decisions are mine.

What that looks like in practice

• The model is never the security boundary. In CareOps and the browser agent, server code decides what the agent can see and do, and the demos show it refusing.
• Every project page says what is live, what is a prototype and what didn't work, with the sources for each fact.
• I run my own infrastructure: a Linux server, Docker, Traefik and Cloudflare DNS.
• Mistakes are recorded. In the agent swarm, review findings go into a ledger, and recurring ones I approve become warnings for the next matching task.

Work I take on

• Designing AI systems and the harness around them: roles, gates, reviews and evaluation.
• Knowledge bases and retrieval that give agents the right context.
• Bounded agents for business workflows, with permissions enforced in code.
• Infrastructure, deployment and full-stack automation.
• Independent testing and evaluation of AI features.