Reasoning over changing infrastructure
Production incidents leave evidence across logs, metrics, traces and deployment history. The research group will train systems to reason over those signals and follow workflows that span multiple tools.
Testing operational decisions
The lab will use simulated environments and synthetic data to evaluate agents before wider deployment. Its remit includes accuracy, reliability and governance for actions in live software environments.
Where the funding goes
- Production-specific models and agent research
- Platform development
- Commercial expansion