vCluster: KubeCon workshop on multi‑tenant LLM infra
vCluster announced a KubeCon EU workshop focused on production‑ready Kubernetes platforms for secure, multi‑tenant LLM/ML workloads with GPU scaling — practical help for teams running shared inference at scale. (x.com)
“Kubernetes Platform Blueprint: From Multitenancy to AI Infrastructure” is a hands‑on workshop scheduled for 11:30 CET on March 23, 2026 at the Van der Valk Hotel Amsterdam Zuidas and requires separate registration from the main KubeCon pass. (vcluster.com) Speakers listed for the session are Daniele Polencic (LearnK8s instructor and Kubernetes author), Artem Lajko (Platform Engineer at iits‑consulting and GitOps author), and Saiyam Pathak (Principal Developer Advocate at vCluster/Loft Labs). (vcluster.com 1) (vcluster.com 2) (vcluster.com 3) vCluster’s product roadmap that feeds into the workshop includes vNode, announced at KubeCon EU 2025 to virtualize Kubernetes nodes and enforce node‑level tenant isolation for workloads such as GPU‑backed inference. (einnews.com) Engineering patterns promised in the workshop map to vCluster features like “Auto Nodes” and dynamic GPU allocation — capabilities vCluster executives described as part of an Infrastructure Tenancy Platform for AI to maximize GPU efficiency in Nvidia Kubernetes environments. (thenewstack.io) (tfir.io) Hands‑on materials are published: the vcluster‑platform‑workshop repository contains exercises for building secure, scalable multi‑tenant Kubernetes platforms with virtual clusters, GitOps, and GPU orchestration. (github.com) Related KubeCon sessions reinforce operational reliability patterns — vCluster’s “Load Testing Argo CD at Scale” documents stress tests up to 1,000 virtual clusters and a €20,000 infrastructure setup for platform validation, and a Mar 24 panel “Running Kubernetes at Enterprise Scale in the Age of AI” includes leaders from JPMorgan Chase and NVIDIA. (vcluster.com) (vcluster.com) vCluster’s upstream footprint and customer signal cited in workshops include a GitHub project with reported usage claims and enterprise references (Adobe, CoreWeave, Atlan, NVIDIA) that platform teams use when assessing production readiness and vendor maturity. (github.com) vCluster is scheduling follow‑up content for operators, including an online livestream “Deploying Managed GPU Clusters for AI at Hyperspeed” on April 14, 2026, for deeper operational patterns around GPU provisioning and multi‑tenant inference. (vcluster.com)
