Design a machine learning solution
Microsoft Learn
Last updated March 16, 2026
A planning-focused path on choosing the right approach for a machine learning project on Azure before any model is built. It works through the decisions that shape a solution: designing a data ingestion strategy by identifying the data source and format and how to serve it, choosing a model training approach, designing a model deployment solution and how deployment requirements feed back into training, and designing a machine learning operations approach for monitoring and retraining in production. The emphasis throughout is on matching requirements to the right Azure service for a given scenario.
What you'll learn
- Designing a data ingestion strategy: source, format, and how to serve data
- Choosing a model training approach for a scenario
- Designing a model deployment solution and how deployment shapes training
- Designing a machine learning operations approach for monitoring and retraining
- Matching project requirements to the right Azure service
Frequently asked questions about Design a machine learning solution
Who is Design a machine learning solution for?
People who need to plan a machine learning solution on Azure and choose the right services and approach.
Is Design a machine learning solution free?
Yes — Design a machine learning solution is completely free to take.
What are the prerequisites for Design a machine learning solution?
None. This is a planning and design path, not hands-on model building; no coding required.
Why we suggest this course
For anyone who has to decide how a machine learning project should be built on Azure — which service, which data strategy, which deployment and operations approach — this path lays out those choices as a structured set of design decisions. One thing to know: it is about identifying requirements and choosing the right approach, not building models hands-on, so it complements build-focused, hands-on training rather than replacing it.