About UTexas
UTexas was founded in 2018 by Dr. Raymond Solis and Dr. Catherine Yue, two professors who had spent their careers studying deep learning systems and distributed computing at leading research universities. They shared a growing frustration: the best ideas in ML research took years to reach practitioners, and when they did, they were often implemented without the statistical rigor that made them valid in the first place. Too many production models were built on intuition dressed up as science.
Raymond and Catherine envisioned a company that operated like a research lab but shipped like a technology firm. They recruited a founding team of PhD-trained engineers and researchers who were equally comfortable writing proofs and writing production code. The first UTexas product — a model evaluation framework grounded in rigorous statistical testing methodology — launched in 2019 and quickly gained a reputation among ML teams for the depth and transparency of its approach.
Today UTexas employs 110 people, roughly a third of whom hold doctoral degrees. We operate a research lab that publishes regularly in top-tier venues like NeurIPS and ICML, a training academy that has graduated over 500 ML professionals, and an engineering organization that builds production frameworks used by 120 organizations. Our offices are in Austin, Boston, and Oxford. We are often asked whether we are a technology company or a research institution — our answer is that we refuse to choose, because we believe the best ML infrastructure is indistinguishable from good science.
Our Mission
To advance the science of machine learning systems through rigorous research, transparent methodology, and frameworks that make empirical evidence actionable.
Our Values
Rigor
Every model we deploy, every claim we make, and every framework we build is held to the standard of peer review. We document our assumptions, disclose our limitations, and publish our methodology so that clients and the broader community can scrutinize and improve upon our work.
Education
Knowledge compounds. We invest deeply in training — both internally and through our public academy — because we believe a better-educated industry produces better models, better systems, and better outcomes for everyone who depends on AI.
Integrity
We will never overfit a benchmark, cherry-pick a result, or sell a framework we do not believe in. Our clients trust us because we tell them what the data actually says, even when the answer is inconvenient. Long-term credibility is worth more than any short-term sale.