In this session, we explored the latest updates in vLLM v0.21.0 with vLLM Core Maintainer Michael Goin, including recent project improvements and ecosystem updates. We then dove into running GenAI workloads without GPUs. The Intel team joined us to share how developers and enterprises can get started with production AI use cases using Intel Xeon CPUs and vLLM, including practical guidance, benchmarks, and deployment considerations. During the Q&A, we discussed CPU-based inference use cases, performance trade-offs, deployment patterns, and where vLLM on Intel CPUs can fit into real-world AI infrastructure. Timestamps: 00:00 Welcome and introductions 04:36 Recent vLLM community content worth checking out 17:00 What’s new in vLLM v0.21.0 23:41 GenAI with vLLM on Intel CPUs 44:38 Live Q&A and discussion Session slides: https://docs.google.com/presentation/d/1i785OasyWQR_OzXocQ8VHd8fnNUJ03vRjTkMQw2LrJE/ Explore and join future vLLM office hours: https://red.ht/office-hours

 550      15