Databricks ML Associate Model Serving Quiz

1. True or False: Batch inference jobs in Databricks can process data from Delta Lake tables.

True

False

2. Which Databricks service integrates with Model Serving to log prediction requests and responses?

Databricks SQL

MLflow and Model Registry

Delta Live Tables

Databricks Jobs

3. In Databricks, what does 'inference' mean?

Training a model on new data

Using a trained model to make predictions

Deleting a model

Storing raw data

4. True or False: MLflow supports model serving only for scikit-learn models.

True

False

5. Which metric is essential to monitor for a deployed classification model?

Disk space usage

Prediction latency and accuracy

Network bandwidth

Cluster memory

6. What is the primary benefit of containerizing models with MLflow?

Reduced model accuracy

Environment reproducibility and portability

Faster training only

Lower storage costs

7. True or False: Model Serving endpoints require a Databricks cluster to run.

True

False

8. How can you version a model in the Databricks Model Registry?

Manual file naming only

Automatic versioning with each registration

Git commits only

Database timestamps

9. Which endpoint type in Databricks Model Serving is optimized for high-throughput inference?

Standard endpoints

Provisioned throughput endpoints

Development endpoints

Batch endpoints

10. What is a key advantage of using Databricks Feature Store in model serving?

It eliminates the need for model training

It ensures consistent features between training and serving

It replaces data warehouses

It only works with SQL queries

11. What is the primary purpose of the Databricks Model Registry?

To store raw training data

To manage model versions, transitions, and lifecycle

To monitor only inference latency

To replace Apache Spark

12. Which Python library is primarily used to log models in Databricks?

TensorFlow only

MLflow

Pandas

NumPy

13. In MLflow, what is a 'registered model'?

A model saved locally on your machine

A model tracked and versioned in the Model Registry

A model only available in notebooks

A model that cannot be deployed

14. What does model monitoring on Databricks help detect?

Data drift and model performance degradation

Only storage capacity issues

Network connectivity problems

User login failures

15. Which authentication method is commonly used for Databricks Model Serving REST endpoints?

Username and password only

API tokens and OAuth

No authentication required

SSH keys only

16. True or False: Databricks Model Serving endpoints automatically scale based on request volume.

True

False

17. What is the MLflow Tracking Server's role in model deployment?

It serves real-time predictions

It logs and stores experiment metadata and models

It replaces Spark clusters

It only handles data versioning

18. Which approach is best for serving predictions on large historical datasets?

Real-time REST API endpoints

Batch inference using Spark jobs

Manual SQL queries

Streaming only

19. In the Model Registry, what does a model stage transition from 'Staging' to 'Production' indicate?

The model has been deleted

The model is ready for deployment in production

The model is still in development

The model requires no monitoring

20. Which Databricks feature enables real-time model serving with low-latency predictions?

Batch inference jobs

Model Serving endpoints

Delta Lake tables

Workflow scheduler

Databricks ML Associate Model Serving Quiz

1. True or False: Batch inference jobs in Databricks can process data from Delta Lake tables.

2.

What first name or nickname would you like us to use?

2. Which Databricks service integrates with Model Serving to log prediction requests and responses?

3. In Databricks, what does 'inference' mean?

4. True or False: MLflow supports model serving only for scikit-learn models.

5. Which metric is essential to monitor for a deployed classification model?

6. What is the primary benefit of containerizing models with MLflow?

7. True or False: Model Serving endpoints require a Databricks cluster to run.

8. How can you version a model in the Databricks Model Registry?

9. Which endpoint type in Databricks Model Serving is optimized for high-throughput inference?

10. What is a key advantage of using Databricks Feature Store in model serving?

11. What is the primary purpose of the Databricks Model Registry?

12. Which Python library is primarily used to log models in Databricks?

13. In MLflow, what is a 'registered model'?

14. What does model monitoring on Databricks help detect?

15. Which authentication method is commonly used for Databricks Model Serving REST endpoints?

16. True or False: Databricks Model Serving endpoints automatically scale based on request volume.

17. What is the MLflow Tracking Server's role in model deployment?

18. Which approach is best for serving predictions on large historical datasets?

19. In the Model Registry, what does a model stage transition from 'Staging' to 'Production' indicate?

20. Which Databricks feature enables real-time model serving with low-latency predictions?