HUD

Builds RL environments and datasets for AI labs.

Builds long-horizon, real-world evals (such as Vending-Bench) by letting AI agents run actual small businesses and studying how they behave.

Research and design lab producing expert-authored task sets for training and benchmarks for evaluation.

Simulation lab producing training data and RL environments so agents can be tested in realistic scenarios.

Open-source computer-use tooling and shared machines where teams and agents work side by side.

Produces custom data for long-horizon reasoning, software engineering and data-science tasks for frontier labs.