An Open-Source Stack for AI Compute - Kubernetes + Ray + PyTorch + vLLM

Explore how to build a comprehensive open-source infrastructure for AI compute through this 16-minute keynote presentation by Ray co-creator Robert Nishihara. Learn about integrating Kubernetes for container orchestration, Ray for distributed computing, PyTorch for machine learning frameworks, and vLLM for large language model serving into a cohesive technology stack. Discover practical approaches to scaling AI workloads using these vendor-neutral open-source tools, understand the synergies between these technologies, and gain insights into building robust AI infrastructure from one of the key architects behind Ray's distributed computing platform.