Managing Thousands of Automatic Machine Learning Experiments with Argo and Katib

Explore the integration of Automated Machine Learning (AutoML) with cloud-native technologies in this conference talk. Learn how to manage thousands of complex hyperparameter tuning experiments using Argo and Katib for optimal performance. Discover best practices, including Argo caching and synchronization, for efficiently developing and deploying AutoML algorithms in production environments. Gain insights into Kubernetes-native workflow orchestration and hyperparameter tuning at scale through practical demonstrations and examples. Understand the architecture of KDP, the benefits of algorithmic workflows, and the implementation of multi-objective optimization. Conclude with a live demo and community discussion, equipping you with valuable knowledge to advance your MLOps capabilities.

Introduction
KDP Overview
KDP Architecture
Why Algo Workflows
Memorization Cache
Template Spec
Example Workflow
Entry Point
MultiObjective Optimization
Implementation
Demo
Community
QA