Prepare for Professional Machine Learning Engineer by connecting data, models, serving, pipelines and monitoring decisions. Follow the guide effective June 1, 2026, and use the worked metrics example to practise choosing the right denominator before interpreting a model result.
Who is the exam for?
Professional Machine Learning Engineer covers building, evaluating, productionizing and optimizing AI solutions on Google Cloud.
The role spans traditional machine learning and generative AI, collaboration, reusable code, data and model pipelines, MLOps and responsible AI. Check the actual objectives for the scope of each topic.
There is no prerequisite certification. Google recommends three or more years of industry experience, including one or more years designing and managing Google Cloud solutions. The general certification FAQ requires candidates to be at least 18.
Official sources: Professional Machine Learning Engineer Certification exam guide (as of June 1, 2026) — Google Cloud Services PDF; Professional ML Engineer Certification | Learn | Google Cloud — cloud.google.com/learn/certification/machine-learning-engineer; Google Cloud Certification FAQs - Cloud Certification Help — support.google.com/cloud-certification/answer/9438208
What does the June 2026 guide cover?
The guide effective June 1, 2026 has six areas with approximate weights of 13%, 16%, 21%, 20%, 18% and 13%.
These published approximate figures total 101%; the table preserves them as printed. Use the linked current guide for complete objectives and product names, including Gemini Enterprise Agent Platform. Check older course coverage against this version.
| Exam section | Approx. weight | What it covers |
|---|---|---|
| 1. Architecting low-code AI solutions | ~13% | BigQuery ML and Agent Platform AutoML for classification, regression, forecasting, clustering; AI APIs and Model Garden models; tuning Gemini, Imagen, Veo; optimizing cost, latency, availability |
| 2. Collaborating within and across teams to manage data and models | ~16% | Data exploration and preprocessing tool choice, feature store, privacy and PII handling, notebook prototyping, experiment tracking, model and lineage tracking, LLM-as-a-judge evaluation |
| 3. Scaling prototypes into ML models | ~21% | Choosing model type, product, and deployment strategy; training data organization; SDKs and custom training; troubleshooting training; hyperparameter tuning; fine-tuning foundation models; CPU/GPU/TPU and distributed training |
| 4. Serving and scaling models | ~20% | Batch and online inference services, containers, model registry and versioning, A/B and canary rollouts, inference pre/post-processing, feature serving, endpoint and hardware choices, scaling the serving backend |
| 5. Automating and orchestrating ML pipelines | ~18% | End-to-end pipelines on managed or unmanaged services, data and model validation, consistent preprocessing between training and serving, retraining policy, CI/CD/CT automation |
| 6. Monitoring AI solutions | ~13% | Security against data exfiltration and malicious prompting, responsible AI and bias monitoring, model explainability, Model Monitoring, training-serving skew, data drift, concept drift, gen AI evaluation |
Official sources: Professional Machine Learning Engineer Certification exam guide (as of June 1, 2026) — Google Cloud Services PDF; Professional ML Engineer Certification | Learn | Google Cloud — cloud.google.com/learn/certification/machine-learning-engineer
Practice precision, recall and accuracy
Precision divides true positives by predicted positives; recall divides true positives by actual positives; accuracy divides all correct predictions by all evaluated items.
The denominators answer different questions. Define the positive class and evaluation set first, then calculate each metric at the stated threshold.
Practice exercise
This fictional evaluation uses one fixed decision threshold and 200 fully labelled items. Defective is the positive class. There are 40 true positives, 10 false positives, 20 false negatives and 130 true negatives. Calculate precision, recall and accuracy. Does accuracy alone establish that the model is suitable for deployment?
Show answer
Precision is 40 / (40 + 10) = 80%. Recall is 40 / (40 + 20) ≈ 66.7%. Accuracy is (40 + 130) / 200 = 85%. Accuracy alone does not establish deployment suitability.
The model flags 50 items as defective, of which 40 are actually defective. It detects 40 of the 60 actual defects and misses 20. The 85% overall accuracy does not convey that recall by itself. The prompt provides no acceptance target, relative error costs or other deployment criteria, so the metrics do not establish a deployment decision or the difficulty of classifying individual items.
Build a study plan using current resources
Map each June 2026 objective to what you can explain, interpret and practise. Use Google’s Machine Learning Engineer learning path, official samples and current technical documentation to fill gaps.
Samples illustrate format and example content; they do not cover the full syllabus or predict a pass. Use an authorized learning environment for applicable hands-on tasks and check access and costs first.
- Low-code solutions: compare the listed model, API and tooling options against a defined task and verify their current capabilities.
- Data and collaboration: review preprocessing, features, privacy, experiments, model tracking and lineage; explain the evidence needed to reproduce a result.
- Training and serving: practise interpreting model, hardware, tuning, inference, scaling and rollout choices against explicit requirements.
- Pipelines: trace data and model validation, preprocessing consistency, automation and retraining objectives for a small learning example.
- Monitoring: solve the classification exercise, then review skew, drift, explainability, responsible AI, security and generative-AI evaluation objectives.
- Before booking: revisit gaps against the current guide and check your appointment requirements.
Official sources: Professional Machine Learning Engineer Certification exam guide (as of June 1, 2026) — Google Cloud Services PDF; Professional ML Engineer Certification | Learn | Google Cloud — cloud.google.com/learn/certification/machine-learning-engineer; Classification: Accuracy, recall, precision, and related metrics | Machine Learning | Google for Developers — developers.google.com/machine-learning/crash-course/classification/accuracy-precision-recall
What should you expect on exam day?
The standard Professional Machine Learning Engineer exam lasts two hours and has 50–60 multiple-choice and multiple-select questions. Registration costs USD200, plus applicable tax. The exam is offered in English and Japanese.
Choose a testing center or Pearson OnVUE remote proctoring. Remote exams are closed book and normally permit no breaks. Check the current ID, workspace, system and appointment instructions; arrange any needed accommodation before scheduling.
Your initial pass/fail result is provisional. Google says to allow 7–10 days for official confirmation. It does not report a numerical score; unsuccessful standard-exam candidates receive section-level feedback to help guide further study.
Professional candidates may make up to four attempts within two years. After a first failure, wait at least 14 days; after a second, 60 days; after a third, 365 days. Attempts count across languages and delivery methods, and each registration is paid.
Official sources: Professional ML Engineer Certification | Learn | Google Cloud — cloud.google.com/learn/certification/machine-learning-engineer; Google Cloud Certification FAQs - Cloud Certification Help — support.google.com/cloud-certification/answer/9438208; Retake Policy - Cloud Certification Help — support.google.com/cloud-certification/answer/9749448; Online Proctored Exams - Cloud Certification Help — support.google.com/cloud-certification/answer/9907748
How does renewal work?
Professional certification is initially valid for two years. The exam renewal window opens 60 days before your inactive date, and exam renewal adds two years to that date.
The current renewal table does not offer this credential a shorter renewal exam or designated Google Skills continuing-education route. Use the standard exam and check your CM Connect profile for eligibility and dates.
Google permits renewal within 30 days after the inactive date, subject to the applicable exam options. More than 30 days later, recertification requires the standard exam at full list price.
Official sources: Professional ML Engineer Certification | Learn | Google Cloud — cloud.google.com/learn/certification/machine-learning-engineer; Google Cloud Certification FAQs - Cloud Certification Help — support.google.com/cloud-certification/answer/9438208; Certification Renewal Options - Cloud Certification Help — support.google.com/cloud-certification/answer/9907853
Official sources
Facts checked against official sources:
- Professional Machine Learning Engineer Certification exam guide (as of June 1, 2026)
- Professional ML Engineer Certification | Learn | Google Cloud
- Google Cloud Certification FAQs - Cloud Certification Help
- Retake Policy - Cloud Certification Help
- Certification Renewal Options - Cloud Certification Help
- Online Proctored Exams - Cloud Certification Help
- Classification: Accuracy, recall, precision, and related metrics | Machine Learning | Google for Developers