About the Exam

Professional Machine Learning Engineer is a Google Cloud certification exam for experienced practitioners who design and operationalize AI and machine learning solutions. It covers low-code AI architectures, collaboration on data and models, scaling prototypes into production models, serving and scaling models, automating ML pipelines, and monitoring AI solutions. Passing demonstrates that you can build and manage scalable ML and generative AI solutions using Google Cloud capabilities and conventional ML approaches.

Exam Topics

Architecting low-code AI solutions13%
Collaborating within and across teams to manage data and models14%
Scaling prototypes into ML models18%
Serving and scaling models20%
Automating and orchestrating ML pipelines22%
Monitoring AI solutions13%

How to Use This Practice Exam

Browse — Read each question, select your answer, and reveal the explanation.
Exam Mode — Simulate real exam conditions with a timed session and score report.
Learn Mode — Spaced repetition schedules questions you struggle with for long-term retention.

Download the Full Exam PDF

Get every question and answer in a clean, printable PDF built for offline study. Purchase once, keep permanent access, and re-download the latest version anytime.

Get PDF

Learn More

Last updated July 26, 2026 at 5:03 AM

Topic filter

Retired questions

Question sort

Questions per page

QuestionQ1

Scaling prototypes into ML models

You are pre-training a large language model on Google Cloud. The model contains custom TensorFlow operations in its training loop. Training will use a large batch size, and you anticipate that it will run for several weeks. You need to configure a training architecture that minimizes both training duration and compute costs. What should you do?

Explanation

Custom TensorFlow operations require a GPU-based training design unless the operations are specifically available and compatible with Cloud TPU; Cloud TPU supports only its documented set of TensorFlow APIs and graph operators. tf.distribute.MultiWorkerMirroredStrategy supports synchronous training across multiple workers with multiple GPUs. Eight a2-megagpu-16g workers provide 128 A100 GPUs while using half as many VM hosts as sixteen a2-highgpu-8g workers, reducing duplicated host resources while retaining the same aggregate GPU count.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!

QuestionQ2

Serving and scaling models

You built a custom model that carries out several memory-intensive preprocessing tasks before making a prediction. You deployed the model to a Vertex AI endpoint and confirmed that results were returned within a reasonable time. After directing user traffic to the endpoint, you find that it does not autoscale as expected when it receives multiple requests. What should you do?

Explanation

Vertex AI scales an endpoint based on CPU utilization by default, with a default target of 60%. Memory-intensive preprocessing may limit concurrent request handling before CPU utilization reaches that target. Reducing the CPU utilization target makes the autoscaler add replicas at a lower utilization level and therefore scale out earlier.

Learn more

Scale inference nodes with autoscaling

Community Discussion

No comments yet. Be the first to start the discussion!

QuestionQ3

Architecting low-code AI solutions

You need to create classification workflows across several structured datasets that are currently stored in BigQuery. Because you will perform the classification multiple times, you want to complete these steps without writing code: exploratory data analysis, feature selection, model building, training, hyperparameter tuning, and serving. What should you do?

Explanation

AutoML Tables provides a no-code workflow for classification on structured data, including BigQuery data. It automates feature engineering, model selection, training, hyperparameter tuning, and deployment, making it suitable for repeated end-to-end classification workflows.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!

QuestionQ4

Monitoring AI solutions

You work for a magazine distributor and must build a model that predicts which customers will renew their subscriptions in the coming year. Using your company’s historical data as the training set, you created a TensorFlow model and deployed it to AI Platform. You need to identify which customer attribute has the greatest predictive influence for each prediction served by the model. What should you do?

Explanation

AI Explanations returns feature-attribution values for each prediction, quantifying how each input feature affected that individual inference relative to a baseline. The Sampled Shapley method approximates Shapley values for the features contributing to the predicted outcome, making it suitable for identifying the most influential customer attribute per served prediction.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!

QuestionQ5

Serving and scaling models

You work for an online travel agency that also sells advertising placements on its website to other companies. You have been asked to predict the most relevant web banner that a user should see next. Security is important to your company. The model-latency requirement is 300ms@p99, the inventory contains thousands of web banners, and your exploratory analysis has shown that navigation context is a good predictor. You want to implement the simplest solution. How should you configure the prediction pipeline?

Explanation

Cloud Bigtable provides scalable, low-latency, high-throughput keyed reads and writes, which suits storing and retrieving each user's navigation context on the online prediction path. An App Engine gateway provides a controlled application layer between the website client and backend services, while managed AI Platform Prediction avoids the additional operational work of hosting the model on Google Kubernetes Engine.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!

QuestionQ6

Automating and orchestrating ML pipelines

You created an ML pipeline that has several input parameters. You want to explore the tradeoffs among different parameter combinations. The parameter options are:

Input dataset
Maximum tree depth of the boosted tree regressor
Optimizer learning rate

You need to compare pipeline performance across the parameter combinations using F1 score, training time, and model complexity. You want the approach to be reproducible and to track every pipeline run on the same platform. What should you do?

Explanation

Vertex AI Pipelines provides a repeatable, parameterized ML workflow, and Vertex AI Experiments can associate multiple pipeline jobs with one experiment so their parameters, artifacts, and metrics can be tracked and compared centrally. Submitting distinct parameter values as separate runs in the same experiment supports reproducible comparison of F1 score, training time, and model complexity.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!

Back home