Kubernetes and Machine Learning

During this workshop you will learn how to run deployments of machine learning (ML) workflows on Kubernetes to make them simple, portable and scalable with open source tools.

Duration: 2 days

$4,000/attendee (minimum 6 attendees)

  • CloudOps Docker and Kubernetes Workshop or good knowledge of these tools.
  • Experience using Linux CLI and a general understanding of virtualization and container technology.
Technical Requirements:

Mac, Linux OS or Windows laptop with SSH client (putty, cygwin), web browser supporting HTML5

Course Outline

Day 1
  • Machine Learning & Kubeflow
  • Deploying Applications with KSonnet
  • Theory & lab: all about KSonnet
  • Kubeflow Introduction
  • Theory about Kubeflow
  • Installation
  • GPUs in Kubernetes
  • Installing Drivers & Docker
  • Deploying NVIDIA GPU device plugin
Day 2
  • Using Tensorflow in Kubernetes
  • TF Training
  • TF Serving
  • TF Batch Prediction
  • Hyperparameter Tuning
  • Other Tensorflow Tools in Kubernetes:
  • JupyterHub
  • TFJobs Dashboard