The fast and the curious: Chasing scalable AI dreams with Kubernetes and k0rdent

Deploying AI at scale on Kubernetes can be a complex and costly endeavor, with challenges like GPU provisioning, cost management, and performance optimization. In this session, Bharath Nallapeta will demonstrate how k0rdent automates GPU-ready cluster provisioning, making AI deployment seamless across cloud and on-prem environments. Attendees will learn how to serve AI models with KServe, dynamically scale GPU resources with Knative auto-scaling, and monitor performance with Prometheus and Grafana. The session will focus on strategies for maximizing compute efficiency and minimizing costs, particularly in a landscape where GPUs are both scarce and expensive. A live demo will walk through the entire AI deployment workflow, from spinning up clusters to running real-time inference, offering insights into streamlining AI operations with Kubernetes without the typical complexity. Learn more: https://platformcon.com/sessions/the-fast-and-the-curious-chasing-scalable-ai-dreams-with-kubernetes-and-k0rdent