About

I'm an applied AI engineer in Brooklyn. I build AI systems that have to work in production, not just in a demo: autonomous agents, content pipelines, and the layer most teams skip — evaluation suites, cost controls, observability, and guardrails. A demo proves your system works once; production proves it keeps working at 2 AM when the model provider starts rate-limiting and users behave in ways you never anticipated.

Some of my personal projects: a trading agent live on Kalshi prediction markets with an hourly learning loop, an Instagram content SaaS, and a taste-gated AI music pipeline. I run these end to end — architecture, code, deployment, and the 2 AM part.

I got here through medical devices. I was lead R&D engineer at Avvio Medical on the Enhanced Lithotripsy System — a device the FDA granted Breakthrough designation: it breaks kidney stones with ultrasound-activated microbubbles in an outpatient visit, no general anesthesia, no stents, no fluoroscopy. I led proof-of-concept through design transfer on its smart catheter, built its stand-alone positioning system as sole contributor, and have patent work to my name from that program. Before Avvio: a mechanical engineering MS at UC Santa Barbara (machine automation), a robotic assembly gantry for CERN's CMS experiment, and hardware for the LUX dark-matter experiment. Physical systems that can't fail taught me how to build software agents that shouldn't.

The fastest way to reach me is the contact form or onofrejauregui@gmail.com. I'm also on LinkedIn and GitHub.