hud

HUD is focused on developing agentic evaluations and reinforcement learning environments for Computer Use Agents (CUAs) that browse the web for advanced AI labs. The CUA Evals framework is the first comprehensive evaluation tool designed for CUAs. It addresses the challenge of determining the reliability of AI agents by providing detailed evaluations across a wide range of tasks. HUD collaborates closely with leading AI labs to deliver scalable agent evaluation and training infrastructure.