We’re a new research lab in San Francisco, currently focused on developing new foundational capabilities for enabling useful AI agents. We’re pursuing several key research bets that will enable AI agents to perform real-world actions, learn from human feedback, self-course-correct, and infer our goals. In particular, we are really excited about the work in combining large language models (LLMs) with reinforcement learning (RL) to solve reasoning and planning, learned world models, and generalizing agents to physical environments. Our work builds on that of Amazon’s broader AGI team, which recently introduced Amazon Nova, a new generation of state-of-the-art foundation models (FMs).

Defining and prioritizing tooling roadmaps that support data collection, model evaluation, and release processes, balancing custom builds. Bridging research and engineering by translating complex technical needs from researchers into actionable product requirements, ensuring tools address real-world bottlenecks and scale workflows.