This is a new site, and a new way I’m working in public.

I’ve spent the last few years shipping AI systems into production at a regulated biotech company. Multi-agent platforms, RAG pipelines, evaluation tooling, the works. Along the way I’ve collected the kind of knowledge that doesn’t make it into papers or framework docs — the specific patterns that work, the failure modes that recur, the architectural decisions you can’t undo.

For a long time, that knowledge stayed in my head and in Team Chats. This site is where it goes from here.

A few things to expect:

Specific over general. I’ll write about real systems and real failures, not “10 tips for RAG.” If a post can’t point to something concrete that happened, I won’t publish it.

Scars over claims. Most of what I know, I learned by breaking things first. I’ll share both the broken version and the fix, because the fix only makes sense once you’ve seen what went wrong.

No hype. The field is loud enough already. I’m trying to add signal, not volume.

If that’s the kind of thing you want to read, you’re in the right place. The first technical post is coming next week.