AI + Kubernetes: What Beginners...
summary
Kubernetes now supports GPU fractionation and driver selection via DRA, in-place pod resize, and gang scheduling so distributed training stops failing when one slice is missing; the repo demonstrates CPU-based simulations, device allocation, distributed training jobs and KServe model serving so beginners can run labs without GPUs. Focus areas: scheduling, hardware allocation, training vs inference, and agent security tradeoffs. Try the Cloud Native AI Lab to follow stepwise labs and reproduce the examples. DRA gang scheduling KServe
AI-written summary · about
chapters (8)
- 0:00AI and Kubernetes intro
- 1:50Kubernetes for AI workloads
- 3:33Scheduling and training
- 5:41Key AI projects on K8s
- 8:14Agnostic AI Foundation
- 10:20Security for AI agents
- 11:46Learning path for beginners
- 13:54AI workload concerns
related talks
38mKubernetes for AI Workloads – What Works and What… Doesn’t
25mExtending Kubernetes for AI | Lessons Learned From Platform Engineering
30mSo You Want To Run AI Agents on Kubernetes: A 101 Guide
38mAI in Kubernetes: How to Get Started?
32mScaling Kubernetes Clusters for Generative Models: Managing GPU Resources for AI App...
30mRunning AI Workloads in Containers and Kubernetes
09mKeynote: Kubernetes for GenAI: The New Operating System of the AI-First World
41mPanel: AI Infra Best Practices: Enterprise Do’s and Don’ts