AI-900 Practice Question: Describe fundamental principles of machine learning on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Uploading training data to Azure Blob Storage for model training

  Why it's wrong here

  Uploading training data to Azure Blob Storage is a data ingestion or preparation activity that occurs before any model is trained; it simply places raw or processed datasets into cloud storage so that training compute can read them. Although Blob Storage can also hold model artifacts, uploading data is not the same as creating an endpoint. Deployment is the later act of exposing a trained model as a callable inference API, independent of where the training data physically resides.

• ✓

  Making a trained model available as a callable endpoint for applications to use

  Why this is correct

  Deployment is the process of taking a trained and registered model and hosting it as an inferencing service—typically a REST endpoint—that applications can invoke with JSON payloads to receive predictions. Azure Machine Learning managed online endpoints handle model versioning, traffic splitting, authentication, and autoscaling for real-time scenarios; batch endpoints serve asynchronous, large-scale scoring. This is exactly what makes a model accessible for real-world use after the training phase is complete.

• ✗

  Distributing the training job across multiple compute nodes

  Why it's wrong here

  Distributing a training job across multiple compute nodes is a parallelization technique used during the model-building phase; Azure Machine Learning splits data or model parameters across a cluster to reduce training time. It does not make the trained model available to applications. Deployment is the subsequent step that packages the trained model into a hosted service, such as a managed online endpoint, for live inference.

• ✗

  Publishing a model to the Azure Marketplace for other organisations to purchase

  Why it's wrong here

  Publishing a model to the Azure Marketplace is a commercial packaging action for offering a prebuilt solution to external customers, not an operational step that lets your own applications call the model. Marketplace listings center on discovering, purchasing, and licensing solutions. In contrast, deployment registers the model in the Azure ML workspace and presents it as a versioned REST endpoint with a scoring script, environment, and inference server for your own apps to consume.

Go deeper

Related to this question