Domain 1 of 5 · Chapter 2 of 3

Machine Learning Workspace Assets

Unlock the complete study guide + 1,040 practice questions across 16 full exams.

Bundled into the existing Operationalizing Machine Learning and Generative AI Solutions premium course — no separate purchase.

14-day money-back guarantee — no questions asked.

Included in this chapter:

  • Assets vs. resources: the reuse model
  • Data assets: three types, immutable versions
  • Environments: curated vs. custom runtimes
  • Components: the reusable pipeline step
  • Registries and asset-level RBAC
  • Exam pattern recognition

Azure ML workspace asset kinds

AspectData assetEnvironmentComponent
What it capturesA named pointer to data plus metadataThe software runtime (packages, Docker image)A pipeline step: code, typed I/O, command, environment
Types / formsuri_file, uri_folder, mltableCurated (read-only) or customCommand component (type: command)
Reference formazureml:<name>:<version>azureml:<name>:<version>azureml:<name>:<version> or @latest
Register withaz ml data createaz ml environment createaz ml component create
Lifecycle verbsVersion, archive, restore (no delete)Version; clone curated to customizeCreate, update, archive, restore
Shared across workspaces viaA registryA registryA registry

Decision tree

Reusable software runtime?packages / Docker imageChange a Microsoft stack?edit a curated environmentReference to stored data?What data shape?Reusable pipeline step?Curated environmentAzureML- prefixCustom environmentconda / Dockeruri_filesingle fileuri_foldera foldermltabletabular / AutoMLComponenttype: commandNot an assetsee resourcesYesNoNoYesYesNofilefoldertableYesNoShare any asset (dev → test → prod, or cross-region): publish to a RegistryAzureML Registry User to use assets; Contributor / Owner to create it
Cheat sheet

Cheat sheet

  • Data assets have three types: uri_file, uri_folder, mltable
  • Data asset versions are immutable; you archive, not delete
  • Jobs mount or download data assets to compute
  • Environments encapsulate the reproducible software runtime
  • Curated versus custom environments
  • Components are the reusable building blocks of pipelines
  • Command components are defined in YAML and registered
  • Registries share and promote assets across workspaces
  • AzureML Registry User role scopes registry access

Unlock with Premium — includes all practice exams and the complete study guide.

References

  1. Create and manage registries - Azure Machine Learning
  2. Data concepts in Azure Machine Learning - Azure Machine Learning
  3. Working with tables in Azure Machine Learning - Azure Machine Learning
  4. Create Data Assets - Azure Machine Learning
  5. Access data in a job - Azure Machine Learning
  6. About Azure Machine Learning environments - Azure Machine Learning
  7. Manage Azure Machine Learning environments with the CLI & SDK (v2) - Azure Machine Learning
  8. What is a component - Azure Machine Learning
  9. Create and Run Component-Based ML Pipelines (CLI) - Azure Machine Learning
  10. Machine Learning registries - Azure Machine Learning
  11. Manage roles in your workspace - Azure Machine Learning