Research
Research that feeds production.
Our research agenda follows the problems we meet in client work, and the results go straight back into it.
World Models
Learned models of an environment that predict how it responds, for simulation and planning.
Distillation
Compressing frontier capability into small models you can own, audit and run yourself.
AI Safety
Evaluation, robustness and human oversight for systems used in high-stakes settings.
Autonomous AI
Agents that plan, act and know when to hand the decision back to a person.
Diffusion Models
Denoising generative models across images, signals and, increasingly, language.
Notes from the practice.
Follow on SubstackLet’s talk
Start a conversationHave a complex AI problem worth solving?
Every engagement starts with a scoping session at no cost. We reply within one business day.