Our lab studies how agents plan, coordinate, and stay safe at scale. We publish openly because better autonomy is a problem worth solving together.
Decomposing open-ended goals into verifiable task graphs that recover from failure.
Protocols for negotiation, delegation, and shared memory across heterogeneous agents.
Alignment, guardrails, and interpretability so agents stay within intent at scale.
We show that uncoordinated agents converge on stable, efficient role divisions when given a shared memory bus and a lightweight market for tasks.
A planning framework where every node carries its own success predicate, cutting silent failures by 71% across 40 benchmark workflows.
An intermediary that intercepts every tool call against declarative policy, enabling safe autonomy without sacrificing throughput.
A logging format that makes any agent run fully reproducible, enabling true debugging and post-hoc audit of autonomous decisions.
An open benchmark of 120 multi-agent tasks with verifiable success criteria.
View on GitHubWe're hiring research scientists and engineers who want to define how autonomous systems work.