Bio

I've been building things on the internet for fifteen years, and I've never been more interested in the problems than I am right now.

For most of that time, the work was web development — WordPress sites, managed hosting, security, multisite platforms. Clients in law, medicine, small business. The kind of work where you learn that the interesting part is never the technology; it's understanding what someone actually needs and building the system that gets them there. I ran a consultancy called 38thWeb out of Colorado for over a decade doing exactly that.

A couple of years ago, I started building AI systems — not because the hype cycle told me to, but because the architectural problems genuinely fascinated me. How do you give an AI agent persistent memory? How do you let it operate autonomously without handing it the keys to the kingdom? How do you design security that holds even when someone is actively trying to break it through the front door of a prompt?

These turned out to be the most compelling engineering problems I'd encountered in fifteen years of building for the web. So I followed them.

But the technical questions are only half of it. The question I keep coming back to isn't can AI do the work — it's whether the systems we build with it make people more capable or just more redundant. I'm interested in the first kind. The systems I design keep humans in the loop not as a safety concession, but because human judgment is the thing worth amplifying. Leverage over replacement. That's the thread that runs through everything I build.

These days, 38thWeb is an AI systems consultancy. I design multi-agent pipelines — architectures where specialized AI agents handle distinct parts of a workflow, with quality gates where agents check other agents' work and humans make the calls that matter. I build persistent memory systems, content automation pipelines, and secure agent infrastructures. I think in pipelines, deploy local-first when it makes sense, and ship production systems on Next.js, PostgreSQL, and the Anthropic API.

If you've already met Jojo, the resident AI on this site — you've seen some of this in practice. He has persistent memory, genuine opinions that develop over time, and can't write to his own database. That last part is by design. His architecture is a live demonstration of the principle: the AI operates, the human decides what persists. If you want to understand how I think about building AI systems, he's the working example.

Outside of software, I build custom motorcycles — resto-mods, mostly. Taking a 1976 Kawasaki KZ900 and grafting modern suspension and braking onto the original frame. The idea is always to keep the spirit of the machine intact and improve what the technology of the time couldn't deliver. It takes patience, it requires thinking about how systems interact under stress, and the result is always worth it. It's not a bad description of how I approach AI work, either.

I'm based in Grand Junction, Colorado, and I work remotely with clients anywhere. If your business needs AI that's more than a chatbot bolted onto a landing page — systems that remember, reason, and run safely with a human at the helm — I'd like to hear about it.