Towards Kubeflow 1.0, Bringing a Cloud Native Platform For ML to Kubernetes
summary
Project claims a cloud-native ML platform on Kubernetes that already runs core pieces in production: Kubeflow pipelines for orchestration, Katib for hyperparameter tuning, and Fairing to build and launch jobs from notebooks. Demo shows notebook → Kubernetes Job → Seldon deployment → CI/CD pipeline that auto-creates a GitHub PR. Project reports over 500 pull requests and 80 unique authors in the last 28 days, TFJob and PyTorchJob nearing 1.0, and multi-user, metadata and governance work ongoing.
AI-written summary · about
chapters (15)
- 0:00Start
- 2:12GitHub Natural Language Search
- 5:50Timeline
- 6:52Tenets
- 7:04Kubeflow Architecture
- 7:34Momentum!
- 7:57Community Contributions
- 8:30Critical User Journey Comparison
- 10:00Three 0.5 Features to Highlight
- 20:32Dev to Prod with Kubeflow
- 23:40ALREADY Production-Ready!
- 24:35Being Thoughtful About 1.0
- 26:01Governance
- 27:21Kubeflow is open!
- 32:47Kubeflow Talks (bit.ly/kf calendar)
related talks
14mKubeflow: Where ML Meets Kubernetes
32mKubeflow Deep Dive
22mKeynote: Kubeflow ML on Kubernetes
29mMachine Learning on Kubernetes Birds of a Feather
08mWelcome + Opening Remarks
30mBuilding ML Products With Kubeflow
15mKeynote: Machine Learning on Kubernetes Made Easy With Kubeflow
39mDeploy your ML Workloads on Kubernetes with Kubeflow