KubeFlow BoF
summary
Room discussion focuses on actionable work: a KFServing standard API is being defined to let one model package deploy to TensorFlow Serving, Seldon, or cloud hosts; a Fairing notebook packaging demo shows notebooks turned into containers and deployed without Kubernetes knowledge; a metadata context prototype captures run ID, params, secrets, artifacts and reads from local or S3 and it works in live demos; migration off ksonnet to kustomize-based installs and a git-triggered Kubeflow pipeline CI/CD demo (2:55 session) are planned.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 0:38Welcome
- 2:28Roadmap for KubeFlow
- 5:50McGraw Co
- 6:25Multitenancy
- 13:31Installation
- 15:43Metadata
- 21:07Summary
- 23:37Data Science Pipeline
- 24:11RealTime Pipeline
- 24:45Monitoring Tools
- 25:12Data Versioning
- 27:25Data Management
- 31:30Parameters
- 32:57Accessing artifacts
- 33:57Injections
- 39:30Visualization
- 40:25ML Spec Discuss
related talks
31mIntroduction to Using and Use Cases of KubeFlow
21mIntroducing Kubeflow (w. Special Guests Tensorflow and Apache Spark)
34mKubeflow: Multi-Tenant, Self-Serve, Accelerated Platform for Practitioners
46mKubeflow, TensorFlow on Kubernetes
24mThe Good, the Bad, and the Missing Parts of Kubeflow
31mKubeflow 2026 State of the Union: Orchestrating the Full ML Lifecycle
14mKubeflow: Where ML Meets Kubernetes
29mKubeflow: TensorFlow on Kubernetes