AWS MLA-C01 vs GCP Professional Machine Learning Engineer: Which ML Certification Should You Get?

AWS and Google Cloud both sell a certification for people who build and operate machine learning systems, but they sit at different levels and test different jobs. The AWS Certified Machine Learning Engineer — Associate (MLA-C01) is an associate-level exam focused on SageMaker and the AWS ML stack. The Google Cloud Professional Machine Learning Engineer (PMLE) is a professional-level exam focused on production ML on Google Cloud (the June 2026 guide moves from Vertex AI naming to Gemini Enterprise Agent Platform), and it assumes more hands-on experience going in.

They are not a like-for-like swap — AWS does not currently publish a professional-level ML engineering exam, and Google does not publish an associate-level one. If you are choosing between the two, the honest framing is “associate AWS vs. professional GCP,” not “same tier, different cloud.” This guide compares both directly against their current official exam guides so you can pick based on your experience level and which cloud your team actually runs on.

Practice this exam in timed mode with the StudyKits Professional Machine Learning Engineer app. All guides and practice apps for these vendors are on the AWS certification hub, Google Cloud certification hub.

The Comparison Table

Factor	AWS MLA-C01	GCP Professional ML Engineer
Level	Associate	Professional
Prerequisites	None formally; AWS recommends 1+ year of hands-on experience with SageMaker and ML workloads on AWS	None formally; Google recommends 3+ years industry experience, including 1+ year designing/managing ML solutions on Google Cloud
Questions	65 (50 scored, 15 unscored)	50-60
Duration	130 minutes	120 minutes
Question types	Multiple choice, multiple response, ordering, matching, case studies	Multiple choice, multiple select
Passing score	720 out of 1000 (scaled)	Not published (Google does not disclose a numeric cutoff)
Exam cost	$150 USD	$200 USD plus applicable tax
Delivery	Testing center or online proctored (Pearson VUE)	Testing center or online proctored (Pearson VUE)
Domain weights published	Yes — four weighted domains	Yes — six sections with approximate weights (guide as of June 1, 2026: ~13/16/21/20/18/13%)

Sources: AWS MLA-C01 exam page and Google Cloud PMLE exam page, checked October 2026 (AWS MLA exam page, Google Cloud PMLE exam page, AWS September 2026 announcement).

Exam update (October 2026): AWS ended English delivery of MLA-C01 on September 28, 2026, and retires MLA-C01 in all languages on January 14, 2027. AWS’s exam page lists MLA-C02 general availability as TBD; AWS’s September 2026 announcement gives January 14, 2027 for MLA-C02 general availability. Google’s PMLE exam guide was revised as of June 1, 2026. The StudyKits MLA app targets MLA-C01, not MLA-C02.

Exam Format Deep Dive

AWS MLA-C01

MLA-C01 runs 130 minutes with 65 questions — 50 that count toward your score and 15 unscored questions AWS uses to evaluate future exam content, mixed in without being labeled. Beyond standard multiple-choice/multiple-response items, AWS’s newer associate and professional exams (MLA-C01 included) use ordering questions, matching questions, and case-study question sets that present a scenario and then ask several independently-scored questions against it.

The exam guide organizes content into four weighted domains: Data Preparation for Machine Learning (28%), ML Model Development (26%), ML Solution Monitoring, Maintenance, and Security (24%), and Deployment and Orchestration of ML Workflows (22%). Data preparation being the single largest domain surprises a lot of candidates who assume the exam is mostly about model training — in practice, close to a third of the questions are about ingesting, cleaning, labeling, and engineering features for ML, not about choosing or tuning algorithms. A passing score is 720 on a 100-1000 scaled range, and the exam costs $150.

GCP Professional Machine Learning Engineer

PMLE runs two hours (120 minutes) with 50 to 60 multiple-choice and multiple-select questions — no ordering, matching, or case-study formats. Google’s exam guide lists six content areas rather than four: architecting low-code AI solutions, collaborating within and across teams to manage data and models, scaling prototypes into ML models, serving and scaling models, automating and orchestrating ML pipelines, and monitoring AI solutions. The guide (as of June 1, 2026) gives approximate weights of ~13%, ~16%, ~21%, ~20%, ~18% and ~13% for these sections. Google does not publish a numeric passing score — results come back as pass/fail only.

Google’s own certification page recommends 3+ years of industry experience with at least one year designing and managing ML solutions specifically on Google Cloud. That is a meaningfully higher experience bar than AWS states for MLA-C01, and it lines up with PMLE sitting at the professional tier rather than associate. The exam costs $200 plus applicable tax and is available in English and Japanese.

Where the Two Overlap

Despite the tier gap, both exams test the same underlying ML engineering lifecycle:

• Data preparation and feature engineering: MLA-C01 dedicates its largest domain to this; PMLE folds it into “scaling prototypes into ML models” and the collaboration domain
• Model training and evaluation: SageMaker training jobs and hyperparameter tuning on AWS; Vertex AI training and AutoML on Google Cloud — different tools, same underlying concepts (bias/variance, evaluation metrics, hyperparameter search)
• Deployment and serving: SageMaker endpoints and batch transform vs. Vertex AI endpoints and batch prediction — both exams expect you to reason about latency, cost, and scaling trade-offs when choosing a serving pattern
• Pipeline automation and MLOps: SageMaker Pipelines and CI/CD on AWS; Vertex AI Pipelines and Kubeflow-based orchestration on Google Cloud
• Monitoring in production: both test model drift detection, retraining triggers, and operational monitoring once a model is live — MLA-C01 pairs this with a dedicated security domain (IAM, encryption, data protection) that PMLE folds into its “monitoring AI solutions” section instead of breaking out separately

If you already work with one cloud’s ML stack, the underlying skills transfer — what changes is the tool names and, for PMLE, the additional generative-AI-agent tooling Google has been layering into recent revisions of the exam guide.

Which One Should You Get First?

Choose AWS MLA-C01 First If:

• Your organization runs primarily on AWS and your ML workloads live in SageMaker
• You have less than a year of hands-on production ML experience — MLA-C01’s associate level and lower prerequisite bar make it a more realistic first ML certification
• You want a credential with clearly weighted domains and a published passing score, so you know exactly where to focus your study time
• You are earlier in your ML career and want the cheaper of the two exams ($150 vs. $200) as an entry point

Read the full AWS MLA-C01 study guide for a domain-by-domain breakdown and an 8-week study plan.

Choose GCP Professional Machine Learning Engineer First If:

• Your organization runs primarily on Google Cloud and your models are trained and served through Vertex AI
• You already have several years of production ML experience and want a credential pitched at that level rather than an entry-point associate exam
• You are comfortable working from approximate section weights and no published passing score
• You want the credential that currently has no AWS equivalent at the same professional tier

Read the full GCP Professional Machine Learning Engineer study guide for the complete domain breakdown and study plan.

If You Are Choosing Between the Two From Scratch

Match the certification to your cloud and your experience level, in that order. If your team’s models run on AWS, or you are newer to production ML and want an associate-level entry point, start with MLA-C01 — it is cheaper, has a lower stated prerequisite, and publishes a passing score. If your team runs on Google Cloud and you already have multiple years of hands-on ML experience, PMLE is the more directly relevant credential, even though it costs more and demands more preparation. If you are cloud-agnostic and simply want the more widely recognized entry point into ML certifications, MLA-C01’s lower barrier to entry makes it the more approachable first step for most candidates.

Preparing with StudyKits

StudyKits offers dedicated practice question sets for both AWS Machine Learning Engineer — Associate and GCP Professional Machine Learning Engineer, each organised by that exam guide’s domains, with offline access and detailed answer explanations on iPhone, iPad, Mac and Android. The MLA app follows the MLA-C01 guide; if you are booking MLA-C02, check the new guide’s changes first.

If your work spans both clouds, MLA-C01 first gives you a structured, lower-cost introduction to certified ML engineering concepts that transfer directly when you tackle PMLE’s broader syllabus afterward.

The Bottom Line

MLA-C01 and PMLE are not competing for the same slot on your resume — one is AWS’s associate-level ML engineering credential, the other is Google Cloud’s professional-level one, and there is currently no direct AWS equivalent to PMLE’s tier. Pick based on which cloud your ML workloads actually run on and how much hands-on production experience you already have, and verify the current domain list on the vendor’s exam guide before committing to a study plan — both providers update these regularly.

Keep reading

• MLA-C01 vs AIF-C01: Which AWS AI Certification First?