Research
Who we are
We are a team of AI researchers and engineers formerly from companies such as Meta AI, Amazon AGI, and Google. Previously, our team has worked on AI research spanning LLM evaluation, fairness, alignment, and embodied agents
Foundational Research for Foundational AI
Deep Research
Understanding and reasoning over large semantic datasets
Multi-Turn Interaction
Enforcing multi-step workflows and supporting dialogue between the user and the agent
Long-Horizon Tasks
Developing agents to take on real-world tasks on the horizon of weeks, months, and even years
Memory
Increasing agentic memory with context windows and other tooling
Research Work
Research Focus Areas
We develop scalable methods to predict, interpret, and steer AI agent behavior, researching the alignment and explainability methods needed to keep increasingly capable systems aligned.
We study how agents think, negotiate, and remember, tackling failure modes like sycophancy and cognitive dissonance while advancing the memory and reward architectures that form the core of robust embodied intelligence.
We pioneer training paradigms that move beyond conventional tool-dependent RL, using evolutionary methods, automated environment design, and world simulation to build agents that learn efficiently and generalize broadly.
Latest Updates
Concrete benchmarks, environments, and methodologies built directly from our research
