Skip to main content

Onboarding

Lifecycle stages:

Introduction
Onboarding
Supported User Workflows
Infrastructure & Compute
Development Environment
Experimentation
Training
Deployment
Monitoring and Alerts

Managed Azure ML helps you get working models into production faster while reducing administrative overhead.

What is Managed Azure ML?​

In Equinor's AI Platform, the Traditional AI/ML journey runs on managed Azure ML. This means that it builds on the capabilities of the Azure Machine Learning platform to provide an opinionated, out-of-the-box setup of resources, permissions, and workflows, so you can focus on building and delivering ML solutions instead of spending time on infrastructure setup, access management, or workflow orchestration. This managed process makes it faster and easier for teams to take machine learning projects from experimentation to production.

What do you get out of the box?​

  1. Single click pre-configured VSCode or JupyterLab IDE.
  2. Pre-configured infrastructure and permissions management with three different tiers to choose from.
  3. Automated CI/CD pipelines using Github Actions.
  4. Template repository to get you started quickly .
note

More information about the CI/CD pipelines and what common ML tasks they can help you with can be found in the supported user workflows section.

Goals of Managed Azure ML​

The key goals of Managed Azure ML are:

  • Accelerate time‑to‑production
    Provide an end‑to‑end set of workflows that support the full ML lifecycle, from experimentation (inner-loop) to deployment across isolated test and production environments (outer-loop).

  • Simplify infrastructure management
    Handle all Azure resource provisioning, access configuration, and environment setup automatically across three standardized infrastructure tiers.

  • Offer a proven, opinionated starting point
    Deliver ready‑to‑use GitHub workflows, repository template, and best‑practice recommendations for local and remote Azure ML development.

  • Support customization
    Allow teams to adjust workflows, repository structures, and deployment approaches if they have advanced or non‑standard needs.

Documents available for this stage​