Red Hat is taking part in a new U.S. National Science Foundation (NSF) research effort aimed at making AI and machine learning infrastructure more efficient. Red Hat Research announced on October 28 that it, IBM Research and other partners will work with the University of California, Berkeley’s Auto-Optimized Agents team in the STRIDE AI Efficiency Challenge (AIEC).
The project
The Berkeley team is building an open source stack that uses real usage traces to automatically optimize deployed coding and enterprise AI agents from end to end. For admins running agents in production, the idea is that the system learns from how agents are actually used rather than relying only on hand tuning.
About the challenge
- The AIEC is supported by the NSF in partnership with the innovation network Start2 Group.
- It includes nine research teams across several universities and technology companies, each eligible for up to $3.5 million per project.
- The goal is to shorten the path from research to real-world enterprise use.
- Teams started work in September 2026, and the full cohort met for its first workshop in San Francisco on October 20.
The news fits Red Hat’s broader push into open source AI infrastructure, including its vLLM and llm-d inference work.
Source: Red Hat Research.