Azure Model Registration and Deployment Quiz

1. In Azure ML, what is model profiling used to determine?

CPU and memory requirements for deployment

Model accuracy on test data

Training time and cost

Data quality issues

2. Which Azure service is used for continuous model retraining and automated MLOps pipelines?

Azure Data Factory

Azure Machine Learning Pipelines

Azure Logic Apps

Azure Functions

3. What is the purpose of model versioning in Azure ML?

Track model changes and enable rollback to previous versions

Compress model file size

Increase model accuracy

Reduce deployment time

4. Which metric is critical to monitor after deploying a model in production?

Model drift and prediction latency

Training accuracy only

Dataset size

Number of features

5. In Azure ML, what does a model's URI or endpoint URL enable?

Direct access to the model for making predictions

Retraining the model automatically

Deleting old model versions

Viewing model source code

6. What is the role of Azure Container Registry (ACR) in model deployment?

Store Docker images for deployed models

Execute training jobs

Manage dataset versions

Monitor model performance

7. Which Python SDK class is used to register a model in Azure ML?

Model.register()

Dataset.register()

Pipeline.register()

Compute.register()

8. When deploying a model, what does a deployment configuration specify?

Compute resources, instance count, and CPU/memory allocation

Model training hyperparameters

Input feature names

Model performance metrics

9. What is the purpose of environment variables in Azure ML deployments?

Configure runtime settings and secrets for the deployed service

Define model architecture layers

Specify training data location

Set model version numbers

10. Which Azure service provides a fully managed MLOps platform for model lifecycle management?

Azure Synapse Analytics

Azure Machine Learning

Azure SQL Database

Azure Data Factory

11. What is the primary purpose of model registration in Azure Machine Learning?

To version and track models for reproducibility and deployment

To automatically train new models

To delete old models from the workspace

To convert models to different formats

12. What does the run_config object specify in Azure ML?

Environment, compute target, and script for training

Model parameters and hyperparameters

Deployment location and endpoint

Data preprocessing steps

13. Which deployment type is best suited for batch predictions on large datasets?

Real-time endpoint

Batch endpoint

Web service

Managed endpoint

14. In Azure ML, what is a managed online endpoint used for?

Real-time predictions with automatic scaling and monitoring

Batch processing large datasets

Training models in parallel

Storing model artifacts

15. What is the purpose of the init() function in a scoring script?

Load the model and initialize resources once when the service starts

Process each prediction request

Train the model on new data

Validate input data format

16. Which scoring script file is required when deploying a model in Azure ML?

Train.py

Score.py

Deploy.py

Config.py

17. What does an inference configuration define in Azure ML deployment?

The compute target and dependencies for the deployed model

The training parameters for the model

The data used for model evaluation

The model's hyperparameters

18. When deploying a model to Azure Kubernetes Service (AKS), which component manages scaling and orchestration?

Docker

Kubernetes

Flask

Conda

19. What is the primary advantage of using Azure Container Instances (ACI) for model deployment?

Lowest cost and fastest startup time for testing

Best for large-scale production workloads

Highest security and compliance

Support for GPU acceleration

20. Which Azure ML object represents a trained model that has been registered?

Pipeline

Model

Dataset

Compute

Azure Data Scientist Model Registration and Deployment Quiz