Speaker: Amy Unruh

Kubeflow is an OSS machine learning stack that runs on Kubernetes. The Kubeflow project is dedicated to making deployments of ML workflows on Kubernetes simple, portable, and scalable. In this workshop, you will learn how to install and use Kubeflow, including Kubeflow Pipelines, to support an end-to-end ML workflow.  

During the workshop, you'll install Kubeflow from scratch, see how to use Kubeflow's multi-user Jupyter notebook servers and other core components, and build and run Kubeflow Pipelines that support full ML workflows, using both the Pipelines UI and its SDK.  In the process, we'll look at how you can use logging, metrics and visualizations, and metadata/artifact tracking, to support ML workflow evaluation and reproducibility.

Kubeflow is an OSS machine learning stack that runs on Kubernetes. The Kubeflow project is dedicated to making deployments of ML workflows on Kubernetes simple, portable, and scalable. In this workshop, you will learn how to install and use Kubeflow, including Kubeflow Pipelines, to support an end-to-end ML workflow.  

During the workshop, you'll install Kubeflow from scratch, see how to use Kubeflow's multi-user Jupyter notebook servers and other core components, and build and run Kubeflow Pipelines that support full ML workflows, using both the Pipelines UI and its SDK.  In the process, we'll look at how you can use logging, metrics and visualizations, and metadata/artifact tracking, to support ML workflow evaluation and reproducibility.

This is your chance to hear from the experts: ML platform builders and longtime users. What makes ML different from other types of applications and why does it require special tooling? How did they get started and what tools would they use if they were starting today? Where is the low-hanging fruit and how do they recommend acquiring those initial quick wins? Join us to learn what you should watch out for in the initial stages, while you're still learning core concepts.

Moderator: Michelle Casbon