Exam prep for
Google Cloud Certified Professional Machine Learning Engineer
Exam set and awarded by Google Cloud On demand
For experienced machine learning practitioners preparing for the 2027 Google Cloud Certified Professional Machine Learning Engineer certification: strengthen your ability to design, deploy, automate and monitor Google Cloud AI solutions at the level assessed by the exam.
Included with Elandi Pro: practice on all 6 exam domains and unlimited mock exams.
Independent exam prep. Elandi is not affiliated with, endorsed by or approved by Google Cloud, and does not award Google Cloud credentials. Our practice questions and mocks are written by Elandi, not taken from the real exam. Exam names are trademarks of their owners. Google Cloud and related certification names are trademarks of Google LLC.
The exam in numbers
The exam assesses the design, development, deployment and monitoring of machine learning and AI solutions using Google Cloud. Coding is not directly assessed. Preparation time of 40–160 hours is a broad planning estimate (no official figure), not a measured candidate average. Google does not publish candidate volume or pass rates; zero values in those fields mean unknown, not zero candidates or a zero pass rate. The mock pass mark is our practice target, not a Google figure. Your mock exams use this learning plan's question count, time limit and pass mark.
- Questions
- 55
- Time limit
- 120min
- Mock pass mark
- 70%
- 39 correct
- Exam fee (paid to Google Cloud)
- US$200
- When you book the exam
- Sitting
- On demand
- Single select (Radio): 4 options
Upcoming sittings
Pick a sitting when you start. Your workspace counts down to it and paces your prep from there.
- Any day On demand Book any day
What Google Cloud Certified Professional Machine Learning Engineer covers
6 modules · 14 units · ~14h of structured prep. Each exam domain with the weight the exam gives it, and the modules that prepare you for it.
- 01Architecting low-code AI solutions 13%
- 1.1 Developing ML models using BigQuery ML or AutoML on Gemini Enterprise Agent Platform
- 1.2 Building AI solutions using Google Cloud AI APIs or foundational models
- 02Collaborating within and across teams to manage data and models 16%
- 2.1 Exploring and preprocessing data for ML
- 2.2 Model prototyping using notebooks (e.g., Gemini Enterprise Agent Platform Workbench and Colab Enterprise)
- 2.3 Tracking and running ML experiments
- 03Scaling prototypes into ML models 20%
- 3.1 Building models given the task considering cost, complexity, latency, and scalability
- 3.2 Training models
- 3.3 Choosing appropriate hardware for training
- 04Serving and scaling models 20%
- 4.1 Serving models
- 4.2 Scaling online model serving
- 05Automating and orchestrating ML pipelines 18%
- 5.1 Developing end-to-end ML pipelines
- 5.2 Automating model retraining
- 06Monitoring AI solutions 13%
- 6.1 Identifying risks to AI solutions
- 6.2 Monitoring, testing, and troubleshooting AI solutions
Your exam workspace counts down and keeps score
Set your exam day once. The workspace paces your prep in four phases, weighs your coverage by what each topic is worth, and shows where the marks are still open after every mock.
- Overview
- All attempts
- Review
- Preview questions
- Questions
- 55
- Time limit
- 120 minutes
- Pass score
- 70% · 39 correct
days until your exam
6 weeks
Thursday, 24 December 2026
- Learn 8+ weeks out
- Mock and repair 8 to 2 weeks out
- Simulate Last 2 weeks
- Consolidate Last 3 days
Now: Mock and repair. Sit a full mock every week or two, then practise the topics you missed.
How ready you are
- Exam-weighted coverage
- 70% of the plan, counted by what each topic is worth
- Full mocks taken
- 2 Prep guides suggest three to five before the exam
- Latest mock
- 66% 4 points below the 70% pass mark
Where the marks are still open
- Serving and scaling models 20% of the exam 54%
- Monitoring AI solutions 13% of the exam 47%
- Automating and orchestrating ML pipelines 18% of the exam 63%
- Scaling prototypes into ML models 20% of the exam 71%
- Architecting low-code AI solutions 13% of the exam 88%
- Collaborating within and across teams to manage data and models 16% of the exam 100%
How the content is written
Write for experienced ML practitioners preparing for a professional Google Cloud certification. Use precise, scenario-driven explanations that make constraints and Google-native trade-offs explicit.
- Worked examples
- Heavy
- Technical jargon
- Heavy
- Maths & formulas
- Light
- Code
- Light
- Diagrams
- Light
Try a sample question
Elandi wrote these in this learning plan's style. Pick an answer to see whether it's right and why. They are not taken from the exam.
A pipeline trains a model and sends its artifact independently to evaluation and deployment. Deployment must occur only after the candidate passes an evaluation policy, including when evaluation completes successfully but reports poor quality. What should the team change? One answer
An Agent Platform Model Monitoring v1 job needs to detect whether current production feature distributions differ from the distribution used to train the deployed model. The training data is available. Which comparison should the team configure? One answer
An application screens prompts for injection and responses for sensitive-data leakage. The security team must tune the filters and thresholds independently for each direction. Which design best fits? One answer
A team maintains reusable offline features in BigQuery and needs low-latency online feature lookups for prediction without maintaining a separate offline feature store. Which design best fits? One answer
Start your Google Cloud Certified Professional Machine Learning Engineer learning plan.
Start preparing for Google Cloud Certified Professional Machine Learning Engineerwith Elandi Pro today.