James Braza

James Braza joins as a Machine Learning Researcher for The Bike Shop. He spent the past nine years in California building technology and is coming to the University of Chicago to research, create, and inspire. He was most recently a research engineer and founding member of technical staff on the learning team at FutureHouse and its spinout Edison Scientific, post- training LLMs and building frontier agents for science. There he coauthored ether0, a state-of-the-art chemistry reasoning model, PaperQA2, the first open-source superhuman scientific literature agent, and Aviary, an open-source framework for training language agents that FutureHouse built its platform atop.

Before that, he was a senior software engineer on Synthego's innovation team for five years, prototyping liquid handling controllers for end-to-end cell therapy manufacturing and creating a gene knockout recommendation engine. He began his career as an automation engineer in SpaceX's Vehicle Engineering department testing high-flow valves under extreme conditions. Data acquisition and control equipment he designed, such as liquid nitrogen heat exchangers, and automated test scripts he wrote have validated hundreds of valves; some flew in SpaceX's first crewed ISS missions.

James grew up in Connecticut, Massachusetts, and Pennsylvania, and has a bachelor's degree in mechanical engineering from the University of Pittsburgh with an economics minor. He believes in holding oneself to a high standard of excellence, solving hard problems by attacking from several angles with open collaboration, and that mistakes are forgotten with quick iteration. James always takes a 'do it yourself' perspective and considers us to be in the era of the great builder.