From Zero to Shared GPUs: Hands-on GPU Scheduling in Kubernetes with HAMi

This hands-on workshop at ContainerDays Singapore 2026 is led by Jimmy Song (VP of Open Source Ecosystem) and Reza Jelveh (Global Market & Solution Architecture at Dynamia AI), both HAMi contributors. It takes you from a bare cluster to a working GPU sharing setup. HAMi is a CNCF Incubation project for heterogeneous GPU virtualization in Kubernetes, supporting 10+ GPU/NPU vendors. Attendees set up HAMi on their laptop, with or without a GPU, and trace how the device plugin intercepts GPU requests, allocates vGPU resources, and integrates with the scheduler. We follow the HAMi architecture and GPU virtualization model (project-hami.io/docs/core-concepts/gpu-virtualization), then explore the DRA integration path. You will deploy HAMi, schedule fractional GPU workloads, and understand how scheduler policies like binpack, spread, and topology awareness change allocation behavior, and how third-party schedulers (KAI Scheduler, Volcano) compose with HAMi's isolation layer. Lab guide: github.com/Project-HAMi/hami-workshop. Bring a laptop with Go, kind, and a GitHub account.