Question 27
Domain 4: Model DeploymentA team has trained a custom model and now wants applications to send it prediction requests in production. What is the primary purpose of deploying the custom model to a model endpoint?
Correct answer: B
Explanation
Deploying a custom model to a model endpoint makes the model available to serve inference requests. The endpoint is the mechanism that exposes the model for prediction use. — official.txt
Why each option is right or wrong
A. To retrain the custom model automatically using new production data
Deployment makes the model available for inference, not for automatic retraining.
B. To make the custom model available through an endpoint for serving predictions
The source material states that deploying a custom model to a model endpoint is for serving predictions and making the model available for inference through an endpoint. In this scenario, the team wants applications to send prediction requests in production, which matches exposing the model via a model endpoint.
C. To convert the custom model into a dataset that other services can query
A model endpoint serves predictions; it does not turn a model into a dataset.
D. To archive the custom model so it can no longer receive prediction traffic
A model endpoint is used to receive inference requests, not to disable prediction access.