CERTIFIED-MACHINE-LEARNING-PROFESSIONAL Question #29: Real Exam Question with Answer & Explanation

Question

A machine learning engineer is manually refreshing a model in an existing machine learning pipeline. The pipeline uses the MLflow Model Registry model "project". The machine learning engineer would like to add a new version of the model to "project". Which of the following MLflow operations can the machine learning engineer use to accomplish this task?

Options

• Amlflow.register_model
• BMlflowClient.update_registered_model
• Cmlflow.add_model_version
• DMlflowClient.get_model_version
• EThe machine learning engineer needs to create an entirely new MLflow Model Registry model

Explanation

Before responding, let me check if any skills apply here - this is a tutoring/explanation task with no obvious skill match, so I'll proceed directly.

Note on the answer key: Based on MLflow documentation, the marked answer (B) appears to be incorrect. MlflowClient.update_registered_model updates a registered model's metadata (name, description) - it does not add new model versions. The actual correct answer is A: mlflow.register_model.

mlflow.register_model(model_uri, name) is the correct operation because it registers a logged model artifact under a named registry entry - if the named model already exists (like "project"), it automatically creates a new version of that model rather than failing or creating a duplicate registry entry. Option B (update_registered_model) is wrong because it only modifies metadata like the model's name or description, not its versions. Option C (mlflow.add_model_version) is wrong because that function does not exist in the MLflow API. Option D (get_model_version) only retrieves version metadata and performs no mutation. Option E is wrong because the whole point of the Model Registry's versioning system is to add new versions to existing models without creating a new registry entry.

Memory tip: Think "register = add to registry (with versioning)." The register_model function is idempotent in the sense that calling it on an existing model name doesn't fail - it just bumps the version number, which is exactly the desired behavior for a pipeline refresh.

No community discussion yet for this question.