Kubeflow - Data Science on Steroids
summary
Kubeflow on Kubernetes is shown deploying Jupyter notebook servers per namespace with GPU access and persistent volumes, composing end-to-end ML pipelines using the Kubeflow DSL and running them with Argo to produce artifacts like trained models and confusion matrices. Installation uses a package manager plus MetalLB to expose a single IP so the stack can run on on-prem or cloud clusters; pipelines, experiment tracking, and CI/CD integration are demonstrated as reproducible.
AI-written summary · about
chapters (6)
- 0:00Kubeflow introduction
- 7:32Cloud-native ML apps
- 15:47Data center workflow
- 21:59Building ML pipelines
- 27:12Advanced pipeline examples
- 30:08Summary
related talks
14mKubeflow: Where ML Meets Kubernetes
35mKubeflow: Data Scientist ready?
30mKubeflow, Your ML Tools on Kubernetes
36mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)
51mMachine Learning using Kubeflow and Kubernetes
36mKubeflow: Tensorflow on Kubernetes
21mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)
46mKubeflow, TensorFlow on Kubernetes