Episode

AzureML Registries - Enabling Better Collaboration and MLOps

with Seth Juarez, Manoj Bableshwar

AI Show

AzureML Registries is a new capability in AzureML that we are announcing at Ignite, October 12th. Registries are organization-wide repositories of Machine Learning assets such as Models, Environments and Components. Registries promote collaboration among data science teams by enabling discovery and reuse of ML models and related assets, improving productivity. Registries enable MLOps across dev, test and prod environments in which you need to promote assets across multiple AzureML workspaces. Today, we'll explore the pain points and challenges with multi-environment MLOps and how Registries can help overcome those.

Join us every other Friday for an AI Show livestream on YouTube.

Chapters

  • 00:00 - Welcome to the AI Show
  • 00:42 - Welcome Manoj
  • 00:58 - What are Registries - Enterprise Machine Learning Lifecycle in the real world
  • 04:45 - Registries in Azure ML
  • 05:15 - Workspaces vs Registries
  • 07:56 - How to train the model
  • 12:09 - Register the model
  • 18:17 - Recap
  • 19:36 - Learn more
  • 20:21 - Wrap

Connect

Developer
Data Engineer
Data Scientist
Azure
Azure Machine Learning