Our Mission
We build software the way critical infrastructure is built — to be trusted, to last, and to never fail quietly, carrying forward traditions of engineering rigor and long-term responsibility into the world of AI.
By pairing modern machine learning with proven classical AI and human judgment, we make powerful AI and expert software systems that are deterministic, verifiable, and safe — without compromising trust or responsibility. Our first product, BengoAI, brings this to the legal profession: hallucination-proof preconsultation AI system that cannot fabricate an answer or recommendation. We’re in currently deep development and exploration, if you’d like early access or to get in touch, please reach out to contact@univalent.cc.
Our Team
We bring nearly 5 decades of combined experience across systems programming, AI/ML, symbolic systems, and cloud software — and we've built software and our career at companies like Apple, Google, and Microsoft.
-
Brandon Pollack
With nearly a decade of experience shaping flagship products at Microsoft and Google - including Windows, Cortana, Google Assistant, Cloud Dataflow, and ChromeOS - Brandon has engineered systems, applications, and services at the forefront of the AI revolution.
Systems Brandon has worked on are used heavily throughout Google for data intensive operations every day, including processing the colossal amount of data required for Machine Learning research and applications.
In addition to founding Univalent, Brandon is also the founder and representative director of Tokyo Rust, a non profit dedicated to educating and fostering the Rust Community in Japan.
Offline, Brandon explores Japan by motorcycle, scuba dives the local seas, and is an avid runner and lifter.
-
Luke Peterson
Luke is a software engineer, system architect and engineering executive with over two decades of experience. After starting his career at Apple, he co-founded Caustic Graphics which was acquired by Imagination Technologies to enable real-time ray tracing on GPUs.
His expertise includes building and managing solid teams, translating business goals into technical vision, and driving towards delivering that vision.
In addition he has extensive knowledge of the patent systems and has been awarded over 40 issued U.S. patents, with over 100 cases internationally and pending. Recently he has been developing tools to enable fundamental research for new approaches to AI.