Section 1
Fundamentals
Preview
Preview: Design and implement an MLOps infrastructure
- 4 of 40 lesson topics
- 1 overview segment
- 3 core concepts
- 2 exam tips
Lesson Topics
- Azure ML Compute Instances
- Initial Workspace Creation
- Workspace Role Permissions
- Azure ML Datastores
Overview
This section establishes the foundational concepts for building and managing machine learning solutions within Azure Machine Learning. It covers the core resources and organizational structures that form the basis of your MLOps infrastructure, ensuring you understand how to set up and manage workspace-scoped assets effectively.
Core Concepts
- **Azure ML Workspace:** The central management boundary for Azure Machine Learning assets and jobs. Supporting Azure resources can exist independently or be provisioned alongside it, but workspace-scoped assets such as data assets, environments, components, jobs, models, and endpoints are created and managed within the workspace.
- **Compute Instances:** Interactive development environments (managed workstations) used for coding, experimentation, and debugging within the workspace. They provide a pre-configured environment with necessary tools and libraries.
- **Azure ML Compute Resources:** Azure Machine Learning jobs can run on configured compute resources such as compute clusters, compute instances for interactive development, serverless compute where supported, or attached supported external compute where applicable.
Exam Tips
- Always consider the workspace as the management boundary for Azure Machine Learning assets and jobs.
- Understand the difference between interactive development on a compute instance and scalable execution on configured job compute.