Getting started with Kubeflow

The popular open source Kubeflow project is one of the best ways to start doing machine learning and AI on top of Kubernetes. However, Kubeflow is a huge project with dozens of large complex components. Skyler Thomas dives into the Kubeflow components and how they interact with Kubernetes. He explores the machine learning lifecycle from model training to model serving.

Kubeflow consists of dozens of components used to train and serve machine learning models, but the first task is to understand the Kubeflow architecture and how it fits into the larger Kubernetes ecosystem. Skyler Thomas details the performance, availability, and security impacts of various Kubeflow deployment options and the decisions you make in how you use Kubeflow. Once you understand the basics, it’s time for you to go hands-on and actually deploy a subset of the Kubeflow components into a running Kubernetes environment using Jupyter Hub and Jupyter notebooks to import your training data and set up prerequisites for creating training environments using TensorFlow (and PyTorch, time permitting). You’ll begin to submit training jobs and generate and serve models via TensorFlow Serving (and Seldon Core, time permitting).