A Year of Democratizing ML With Kubernetes & Kubeflow
summary
Project hits 0.3 with tooling that actually works: KFctl one-line install and a GCP click-to-deploy that provisions clusters and apps, a Pipelines Python SDK that compiles to YAML and a GUI, autoscaling that spins up GPU node pools for jobs then tears them down, and a StudyJob CRD (Katib) for hyperparameter search showing runs from ~10% to >99% validation accuracy. Community totals ~1,900 commits, ~100 contributors from ~30 companies; deployment and demos run end-to-end.
AI-written summary · about
chapters (9)
- 0:00Start
- 5:01User Experience
- 5:49Multi-Cloud is the Reality
- 6:19Kubecon 2017
- 7:41Momentum!
- 7:57Community Contributions
- 12:31Click to Deploy
- 22:40Autoscaling Jobs
- 31:04We're just getting started! Our roadmap
related talks
37mTowards Kubeflow 1.0, Bringing a Cloud Native Platform For ML to Kubernetes
30mBuilding ML Products With Kubeflow
32mKubeflow Deep Dive
51mMachine Learning using Kubeflow and Kubernetes
22mKeynote: Kubeflow ML on Kubernetes
39mDeploy your ML Workloads on Kubernetes with Kubeflow
15mKeynote: Machine Learning on Kubernetes Made Easy With Kubeflow
46mDeploying scalable Machine Learning models with Kubernetes