Stephen Watt, Red Hat | Red Hat Summit 2025
summary
Claims that Red Hat democratizes fine-tuning with InstructLab and boosts inference performance via its Neural Magic work integrated with vLLM, while contributing upstream to Llama Stack and a model context protocol to enable RAG/RAFT workflows and service calls. Practical tooling includes vector databases for ETL chunking, Red Hat Edge Manager and Advanced Cluster Manager to decide what inferencing runs at the far edge versus mothership, a Kepler project for power-aware scheduling, and an experimental semantic router for model selection; net‑zero by 2030 is the stated sustainability goal.
AI-written summary · about
chapters (5)
- 0:00Intro
- 3:23Advancements in AI: The Role of Open Context and Open Source
- 6:41Empowering AI Innovation: The Role of Open Communities and Red Hat
- 11:44Integrating AI with Edge Computing: Strategies and Real-World Applications
- 15:24Innovations and Commitments: Red Hat's Journey in AI, Sustainability, and Developer Empowerment
