PD Disaggregation VLLM Deployment on Alternative AI Accelerators Using Llm-d - 纪飞 王 & Mengxuan Li
CNCF — Sep 24, 2026 — conference
Why watch
This session explores deploying vLLM with PD (Prefetching-Decoding) disaggregation on alternative AI accelerators using the llm-d project. It targets ML platform engineers and cloud infrastructure developers interested in running LLM inference workloads on diverse hardware within Kubernetes environments, moving beyond NVIDIA-centric deployments.