Azure Machine Learning
Azure Machine Learning (Azure ML) is a cloud service for the ML project lifecycle: training, deploying and managing models, and running MLOps. You can build models in it or bring models built in PyTorch, TensorFlow, scikit-learn, XGBoost or LightGBM (R and .NET also supported). Microsoft Learn page dated 2026-10-19.
Capabilities (Microsoft Learn)
- Interfaces: Azure ML studio (web), Python SDK v2, Azure CLI v2 and Azure Resource Manager REST APIs.
- Studio tools: managed Jupyter notebooks, run-metric visualisation, a no-code designer, an automated ML UI and data labelling.
- Training: cloud scripts, AutoML, hyperparameter tuning, multinode distributed training (PyTorch, TensorFlow, MPI) on compute clusters, serverless compute or Azure ML Kubernetes; Spark through serverless Spark or an attached Synapse pool.
- Deployment: managed endpoints for real-time (online) and batch scoring; online traffic can be split across deployments to test new model versions.
- MLOps: Git and MLflow integration, pipeline scheduling, Event Grid triggers, CI/CD through GitHub Actions or Azure DevOps, job artifacts and lineage between jobs and assets.
- Generative AI: a model catalog (hundreds of models from providers including Azure OpenAI, Mistral, Meta, Cohere, NVIDIA and Hugging Face) and prompt flow.
- Security: virtual networks, Key Vault, Container Registry behind a virtual network, role-based access; data is not stored or processed outside the deployment region.
Prompt flow is being retired
Microsoft states that prompt flow in Microsoft Foundry and Azure Machine Learning will be retired on 2027-04-20, is not recommended for new development, and should be migrated to Microsoft Agent Framework.
Where it fits
Azure ML is the classic-ML and MLOps service in microsoft-azure; generative-AI work is shifting to Microsoft Foundry. It supports mlflow for tracking, Docker-based environments, and Kubernetes through Azure Arc or Azure ML Kubernetes. For model serving at the library level see vllm. Billing is by the Azure resources used; see Microsoft’s pricing page.
Sources
- Microsoft Learn: What is Azure Machine Learning? (2026-11-19)