A data scientist sends a POST request to a Databricks Model Serving endpoint with a JSON payload containing feature values. What does the response typically contain?

The model's predicted output, such as a class label or numeric value

The full training dataset used to build the model

The model's training loss curve

The model's feature importance scores

A model serving endpoint is serving version 3 of a model. The team has trained version 4 and wants to gradually roll it out to a small percentage of traffic before fully switching over. What is the best way to achieve this?

Create a new endpoint for version 4 and update the application to send a percentage of traffic to it.

Update the existing endpoint to serve version 4 and immediately switch all traffic.

Delete the existing endpoint and recreate it with version 4.

Use the endpoint's traffic routing feature to send a small percentage of traffic to version 4.

A machine learning engineer wants to expose a registered model for realtime inference using Databricks Model Serving. Which action is required to create a serving endpoint?

Deploy the model to a running cluster and expose its driver port

Create a job that runs the model every minute and stores predictions in a table

Register the model in Unity Catalog and create a serving endpoint from the Serving UI or API

Package the model as a Docker container and push it to Azure Container Registry

A model serving endpoint is returning incorrect predictions for a small subset of inputs. The model was working correctly in training. The team has verified the input data is correct. What is the most likely cause and the best next step?

The model is overfitting; retrain the model with more data.

The feature encoding in the serving endpoint does not match the training pipeline.

The model is experiencing data drift; retrain the model with recent data.

The endpoint is using a different model version than expected.

A model serving endpoint is returning 5xx errors for a small percentage of requests. The endpoint is configured with autoscaling. The team notices that the errors correlate with times when the CPU utilization of the serving instances is above 90%. What is the most likely cause and the best first step to resolve it?

The model is returning errors for certain input data; validate the input schema.

The endpoint is under-provisioned; adjust the autoscaling configuration to scale out more aggressively.

The endpoint's model version is corrupted; redeploy the model.

The endpoint is hitting a timeout due to high load; increase the timeout setting.

DatabricksCertified Machine Learning Associate

Domain 4Objective 5

Deploy and Query a Model for Realtime Inference MACHINE-LEARNING-ASSOCIATE Practice Questions (Page 3)

Part of the Section 4: Model Deployment domain, which makes up ~30% of our current practice bank. Databricks does not publish an official question count, but from its 90-minute exam (~35–60 total, ~11–18 in this domain), expect 2–3 from this objective — we provide 18 practice questions to prepare you well beyond it. (estimate)

18questions here

4free pages

4concepts

Questions 11–15

11foundation · easy
A data scientist sends a POST request to a Databricks Model Serving endpoint with a JSON payload containing feature values. What does the response typically contain?
Select an answer first
12application · medium
A model serving endpoint is serving version 3 of a model. The team has trained version 4 and wants to gradually roll it out to a small percentage of traffic before fully switching over. What is the best way to achieve this?
Select an answer first
13foundation · easy
A machine learning engineer wants to expose a registered model for realtime inference using Databricks Model Serving. Which action is required to create a serving endpoint?
Select an answer first
14expert · hard
A model serving endpoint is returning incorrect predictions for a small subset of inputs. The model was working correctly in training. The team has verified the input data is correct. What is the most likely cause and the best next step?
Select an answer first
15expert · hard
A model serving endpoint is returning 5xx errors for a small percentage of requests. The endpoint is configured with autoscaling. The team notices that the errors correlate with times when the CPU utilization of the serving instances is above 90%. What is the most likely cause and the best first step to resolve it?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to MACHINE-LEARNING-ASSOCIATE

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Databricks. “MACHINE-LEARNING-ASSOCIATE” is a trademark of its owner, used for identification only.