Galois Research builds intelligent systems for the physical world. We work with organizations whose hardest problems live in rooms, on benches, and on production floors, where the ground truth is physical and being wrong is expensive.
Fields
The internet taught models to read, write, and draw. It did not teach them what a production line does when a part arrives a millimetre out of place, or what a laboratory looks like at the end of a twelve-hour run. That knowledge lives in places cameras rarely enter, and general-purpose models fail there at exactly the moments that matter: the hidden object, the rare failure, the case nobody labeled.
Every one of these fields needs software that understands its own world. That is what we build.
Practice
We take a problem from the first observation to a system in production. We learn from what the environment can show us and build what it cannot, and we judge every result against the conditions that matter rather than an average that hides the failures.
Nothing we deliver rests on a promise. It rests on evidence gathered in the client’s own environment.
Group
Galois Research is a small group of engineers and researchers. We take on a few engagements at a time and work inside the client’s environment, with their data, their constraints, and their people. Our engagements are confidential. We do not name our clients, and we do not describe their work.
If you have a hard problem where software meets the physical world, we would like to hear from you.