Research on AI agent simulation, reliability, and production monitoring.

Evidence-led writing for teams testing long-running agents, comparing complete trajectories, and improving autonomous workflows before and after deployment.