Timed, full-length practice sessions of original PMLE questions, written by Courseiva to the published Google Cloud objectives and spread across all 7 official exam domains.
Practise under timed conditions
To practise under timed conditions, start a 60 or 120-question session, put away all notes, set a timer matching the real exam duration (120 minutes), and commit to each answer before moving forward. This builds the time management and decision-making you need on exam day.
This free PMLE mock exam uses original Courseiva questions, written to the objectives Google Cloud publishes for the Google Professional Machine Learning Engineer exam. Each session spreads questions across all 7 official exam domains in line with the published weightings. It is not the real exam and contains no real exam questions.
PMLE Domain Distribution
Automating and Orchestrating ML Pipelines
Serving and Scaling Models
Architecting Low-Code ML Solutions
Scaling Prototypes into ML Models
Collaborating Within and Across Teams to Manage Data and Models
Monitoring ML Solutions
Collaborating to manage data and models
Every question is checked against the 2026 PMLE exam objectives and published under the editorial oversight of an engineer with 12+ years' experience. These are original practice questions — not dumps — so you build real understanding rather than memorising answers.
Both the mock exam and practice test use the same question bank. The difference is in how you use them — and when to use each during your PMLE study plan.
Practice test — for learning
Use the PMLE practice test when you are studying a domain. Answer questions, read every explanation immediately, and build understanding. Do 10–30 questions per domain per session. This is your primary study tool for the first 4 weeks.
Go to practice test →Mock exam — for simulation
Use the PMLE mock exam in the final 1–2 weeks before your test date. Complete a 60 or 120-question session without stopping, manage your time, then review all results at the end. This builds exam-day stamina and surfaces final weak spots.
Start 120-question mock →Try these sample questions from the mock exam bank. Commit to an answer before revealing the explanation.
A machine learning engineer is building a Vertex AI pipeline that uses a pre-built Google Cloud Pipeline Components (GCPC) to train a custom model. Which component should the engineer use to submit a custom training job to Vertex AI?
Select an answer to reveal the explanation
A team has a Vertex AI pipeline that includes a container component for data preprocessing. The team notices that the component is re-executed every time the pipeline runs, even when the inputs and code haven't changed. They want to leverage pipeline caching to avoid redundant executions. What should they do to enable caching for this component?
Select an answer to reveal the explanation
A data scientist wants to deploy a trained TensorFlow model to Vertex AI for online predictions. They need to serve predictions with low latency and want to leverage GPU acceleration. Which machine type should they select when creating the Vertex AI endpoint?
Select an answer to reveal the explanation
You need to serve a large embedding model for similarity search with low latency. The model was trained to generate 256-dimensional embeddings. You plan to use Vertex AI Vector Search. Which index type should you choose to balance accuracy and performance for a dataset with 10 million vectors?
Select an answer to reveal the explanation
A machine learning engineer needs to run batch predictions on 50 TB of data stored in BigQuery using a Vertex AI model. The model is a custom container. What is the most efficient way to set up the batch prediction job?
Select an answer to reveal the explanation
A retail company wants to predict customer churn using historical purchase data stored in BigQuery. The data includes customer demographics, transaction history, and support interactions. The team is comfortable writing SQL and wants to avoid moving data to a separate environment. Which approach should they take?
Select an answer to reveal the explanation
You have a TensorFlow training script that runs on a single machine. To speed up training on Vertex AI with 8 GPUs on a single machine, which strategy should you use?
Select an answer to reveal the explanation
A data science team is building a feature engineering pipeline that processes large-scale data from BigQuery daily. They need to compute aggregate features and store the results in Vertex AI Feature Store for both online serving and offline training. Which Google Cloud service is best suited for this batch computation?
Select an answer to reveal the explanation
You are fine-tuning a large language model (LLM) from Hugging Face Transformers using Vertex AI Training. The model has 7 billion parameters and does not fit into the memory of a single GPU. You need to train across multiple GPUs, splitting the model layers across devices. Which distributed training approach should you use?
Select an answer to reveal the explanation
A machine learning team wants to share features across multiple models to reduce training-serving skew and ensure consistency. Which Vertex AI service should they use?
Select an answer to reveal the explanation
A data scientist has deployed a model on Vertex AI Endpoints and wants to monitor the model's predictions for any drift over time. Which Vertex AI service should they use?
Select an answer to reveal the explanation
A healthcare organization is building a machine learning model to predict patient readmission risk. They have sensitive data stored in BigQuery that includes protected health information (PHI). The data science team uses Vertex AI Workbench notebooks to explore the data and develop models. The organization's security policy requires that all PHI data must be encrypted at rest and in transit, and that access to the data is logged and audited. They also need to ensure that the data used for model training is de-identified to remove direct identifiers such as patient names and SSNs. The team wants to automate the de-identification process as part of the data pipeline. Which approach meets these requirements?
Select an answer to reveal the explanation
Answer all 12 questions to see your domain score breakdown
Sitting the PMLE under real exam conditions is a skill in itself. Candidates who underperform often do so not because of knowledge gaps, but because of poor time management or test anxiety. Use your final mock exam sessions to address both.
The PMLE exam lasts 120 minutes. Do not spend more than 90 seconds on any single question on the first pass. Flag difficult ones and return to them after completing the rest.
On every question, immediately eliminate obviously wrong choices. Even if you are unsure between two options, narrowing to two doubles your odds. Most PMLE distractors contain a subtle error — re-read the scenario constraint before committing to the answer that sounds most familiar.
Google Cloud writes many PMLE questions as realistic scenarios. Read the final sentence first — it tells you what is being asked. Then re-read the scenario with the question in mind to avoid wasting time on irrelevant details.
The real PMLE is a long, focused sitting (120 minutes). In the week before your exam, complete at least two full timed mock sessions on separate days to build concentration stamina.
Time limit
120 min
Official exam duration
Yes. Courseiva provides free PMLE mock exam questions across all official exam domains. The platform includes timed simulation, per-domain score breakdown, missed-question review, and readiness tracking. No account required — free forever, supported by advertising.
The practice test is optimised for learning: you see explanations after each question immediately. The mock exam is optimised for simulation: you answer all questions under time pressure and review at the end. Use practice tests for studying and mock exams for benchmarking.
Aim for consistent scores of 80% or above on full-length PMLE mock exams before booking your test date.
Most candidates who pass PMLE on their first attempt complete 3–5 full-length mock exams in the two weeks before their test. This is enough to identify final weak spots, build stamina, and verify readiness without over-stressing or running out of fresh questions.
No — all Courseiva questions are original, AI-assisted and checked against the public Google Cloud exam blueprints, with editorial oversight from an experienced network and security engineer. Exam dumps are memorised real exam questions shared illegally. Using dumps violates your Google Cloud certification agreement and can result in your certification being revoked. Our questions make you genuinely competent, not just test-day lucky.
Track your mock exam scores, see per-domain analytics, and benchmark readiness across every certification.
Sign Up FreeFree forever · Every certification included