How Yu Misaki Became a Golden Kubestronaut @cncf

Uploaded August 2026 | Updated September 2026, 3 weeks ago

Hands-on practice and consistent studying took Yu Misaki from zero Kubernetes experience to Golden Kubestronaut. He shared his story at KubeCon + CloudNativeCon Japan.

#KubeCon #CloudNative

How Yu Misaki Became a Golden Kubestronaut

Open Source SecurityCon | Welcome + Opening Remarks - John Kjell & Marina Moore

Score-Driven Multi-Cluster Management: An Evaluation Framework for… K. Takeuchi & J. Banerjee

Sealed for All: Multi-Party Confidential Computing, No Trust - Ryota Hashimoto & Takahiro Kambara

1 Year, 15 Certs, a Raspberry Pi Home Lab: How I Became a Golden Kubestronaut - Yu Misaki

CNCF CTO Chris Aniszczyk: Its Never Too Early to Get Involved

CNCF On-Demand: Live Migration with KubeVirt and Calico

Keynote: Plug in and Scale: Serving LLM Models on Kubernetes Ma... Shrinidhi Venkataraman & Nithin R

Beyond VLLM: Distributed LLM Inferencing With Llm-d on Kubernetes - Ravindra Patil, Red Hat

SBOMit: Making SBOMs Accurate With Attestations - Marco De Vincenzi & Justin Cappos, NYU

Project Lightning Talk: Opening Remarks - Hoon Jo, KubernetesLab.dev

$Beyond Single-Cluster Limits: Scaling GPU Workloads Across Kubernetes With… K. Das & E. Bayramova Dont miss out! Join us at our next KubeCon + CloudNativeCon events in Shanghai, China (8-9 September, 2026) and Salt Lake City, United States (Nov 9–12, 2026). Connect with our current graduated, incubating, and sandbox projects as the community gathers to further the education and advancement of cloud native computing. Learn more at https://kubecon.io Beyond Single-Cluster Limits: Scaling GPU Workloads Across Kubernetes With Virtual Nodes - Kunal Das, Cast AI & Esmira Bayramova, Kimchi Our ML platform team hit a wall when GPU demand became unpredictable. Fine-tuning jobs queued for hours on busy days while expensive on-prem GPUs sat idle on others. Adding cloud GPU clusters solved capacity but fractured our workflow , developers juggled multiple kubeconfigs, Kueue couldnt see cross-cluster queues,workloads couldnt failover because multi-cluster was really just isolated clusters with shared dashboards. In this talk, we walk through how we used Liqos virtual node pattern to unify 5 heterogeneous clusters , mixing on-prem A100/ H100 nodes with cloud spot GPU instances from cloud providers , into a single schedulable topology. Well share how the vanilla Kubernetes scheduler handles placement using standard node selectors and affinities, how Ciliums cluster mesh compared to Liqos WireGuard-based network fabric for cross-cluster pod traffic, and how we integrated Kueue for unified job queuing across the virtual cluster.$

CNCF [Cloud Native Computing Foundation] |

How Yu Misaki Became a Golden Kubestronaut