-A team is standardizing MLOps practices by using automated deployments.

Deploy resources from a pipeline: Azure CLI Commands

Define Azure resources declaratively: Bicep templates

Azure CLI Commands should be used to deploy resources from an automation pipeline. Azure Machine Learning supports integration with CI/CD platforms such as Azure DevOps and GitHub Actions, and Microsoft documents the Azure CLI with the Machine Learning extension as a standard mechanism for automating resource provisioning, training pipelines, model deployment, and other MLOps operations. Azure CLI commands can be executed non-interactively within pipeline stages, making them appropriate for repeatable automated deployments.

Bicep templates should be used to define Azure infrastructure declaratively. Bicep is Microsoft ' s domain-specific Infrastructure-as-Code language for Azure Resource Manager. Instead of specifying individual imperative provisioning steps, a Bicep file describes the desired state of Azure resources. Azure Resource Manager then determines the deployment operations required to reach that state. This provides repeatable, version-controlled, consistent infrastructure across development, testing, and production environments.

Bicep deployments can also be invoked directly through Azure CLI commands such as az deployment group create, allowing the declarative infrastructure definition and automated deployment mechanism to work together in an MLOps CI/CD pipeline.

Study Guide Reference: Design and implement an MLOps infrastructure — Infrastructure as Code, Bicep, Azure CLI, automated deployment pipelines, and reproducible environment provisioning.