Business development for a federal contractor — and the AI tools underneath it, built by hand.
Not an engineer. I ship into government compliance, one of the least forgiving places software can go. Here's the work, and what it's teaching me.
I've spent my career on the business side of federal programs — acquisition, program management, the unglamorous machinery that makes government actually deliver. For most of it, building software meant waiting on someone else.
That changed. AI now lets a domain expert build the tools they used to wait on a dev team for — and that quietly rewrites who gets to build software at all. I'm a test case for the idea, documenting it in the open.
A pre-award-to-post-award platform with 20 specialized AI agents that draft, review, and manage government contracts — checked live against FAR, DFARS, VAAR, and HHSAR rule sets. The work a team of analysts grinds through by hand, structured and accelerated.
A training simulator that coaches government acquisition staff through real procurement decisions — and never hands them the answer. It makes them reason their way to it. Most AI races to answer faster; this one is built to make you better without it.
I run my work and my life on a heavily customized personal AI infrastructure — its own agents, voice, persistent memory, and automated workflows wired into how I actually operate. Built in the open, used every day. My clearest proof the operator-builder thing is real.
Also shipped: interax.app (a drug-interaction predictor, on the App Store), a weather-prediction app, and a pile of tools my kids actually use.
I'm starting to write about what it actually takes to build AI software when you came up on the business side — what works, what's hype, and what shipping into a high-stakes domain demands that the demos never mention. New posts land on LinkedIn first.
Building something in government, federal contracting, or AI tooling — or just want to trade what's working? Based in Málaga, Spain, and easy to find.