What is a benefit of deploying a model using a custom container?

It automatically scales to zero when not in use

It eliminates the need for a model artifact

It guarantees lower latency than managed serving

It allows you to include custom dependencies and libraries not available in standard serving images

What is the primary purpose of a feature store in a machine learning platform?

To store trained model artifacts and manage their versions

To provide a compute engine for training large-scale models

To execute feature transformations in real time on streaming data

To centralize feature definitions, storage, and serving so the same features are used for both training and inference

How does a feature store typically serve online features to a model at inference time?

By running batch MapReduce jobs on the feature data before each prediction

By exposing a low-latency API that retrieves feature values for a given entity key

By requiring the client to compute features and send them with the request

By embedding all features directly into the model weights

A team deploys a large BERT model for sentiment analysis on Vertex AI Endpoints. The model's p99 latency is 800ms, but the SLA requires 300ms. The team has budget for hardware changes but wants to minimize cost. What should they do first?

Increase the number of replicas to spread the load

Use a larger machine type with more vCPUs

Quantize the model to int8 and measure the latency impact

Switch from a CPU machine to an A100 GPU

Google CloudProfessional Machine Learning Engineer

Domain 4Objective 2

Scaling Online Model Serving PROFESSIONAL-MACHINE-LEARNING-ENGINEER Practice Questions (Page 2)

Part of the Serving and scaling models domain, which accounts for ~20% of the PROFESSIONAL-MACHINE-LEARNING-ENGINEER exam.

37questions here

8free pages

14concepts

~20%of the exam

14 Concepts in This Objective

Feature Store fundamentals
Serving features from Feature Store
Feature Store configuration and monitoring
Public and private endpoints
Deploying models to endpoints
Endpoint security and access control
Hardware selection for serving
Hardware optimization and cost trade-offs
Scaling serving infrastructure
Containerized serving and inference platforms
Throughput and latency management
Model tuning for production
Serving-specific model optimization
Production model lifecycle management

Questions 6–10

6foundation · easy
What is a benefit of deploying a model using a custom container?
Select an answer first
7foundation · easy
Which metric is most directly used to measure the responsiveness of a model serving endpoint?
Select an answer first
8foundation · easy
What is the primary purpose of a feature store in a machine learning platform?
Select an answer first
9foundation · easy
How does a feature store typically serve online features to a model at inference time?
Select an answer first
10application · medium
A team deploys a large BERT model for sentiment analysis on Vertex AI Endpoints. The model's p99 latency is 800ms, but the SLA requires 300ms. The team has budget for hardware changes but wants to minimize cost. What should they do first?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to PROFESSIONAL-MACHINE-LEARNING-ENGINEER

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 Google Cloud. “PROFESSIONAL-MACHINE-LEARNING-ENGINEER” is a trademark of its owner, used for identification only.