Jaime Gutierrez

I've spent my career on the production side of scientific data. I started as an end user in a genetics lab fixing the systems I depended on because they were broken, and spent my career taking labs from Excel to LIMS systems.

The thing I care about is data that stays trustworthy as it scales: strong schemas and clear ownership, so nobody has to take a record on faith. That instinct is what I now apply to AI agents.

I run extensive planning sessions with agents up front, turn the plan into a declarative dependency graph of work items, and let a merge gate enforce ordering and quality from that graph. Agents do more, faster, and the rules they operate under are structural, not hopeful prose.

Currently building tapir-os, a Rust-based system for running agent fleets under a single operator on exactly those principles.