AWS Certified Machine Learning Engineer Associate: Exam Cost, Domains and the MLA-C02 Switch

The AWS Certified Machine Learning Engineer Associate tests whether you can prepare data, build, deploy and monitor machine learning models on AWS. AWS is switching it from exam MLA-C01 ($150 USD, 65 questions, 130 minutes) to an updated MLA-C02, which is in beta. It’s built for people with about a year of ML and AWS experience.

What Is the AWS Certified Machine Learning Engineer Associate?

It’s an associate-level exam from Amazon Web Services for people who work, or want to work, on the engineering side of machine learning. The focus is on the pipeline around a model: getting data ready, training and tuning, deploying, and keeping the model healthy once it’s live.

That’s different from the AWS Certified AI Practitioner, which is foundational and tests concepts. It’s also different from a research-style credential. This one asks whether you can operate machine learning on AWS.

Who it fits:

• Software engineers and data engineers moving into machine learning work.
• Machine learning practitioners who want AWS credentials.
• AI degree students who already use AWS in their coursework or an internship.
• Cloud engineers who are now responsible for model deployment.

If you’re brand new to AI, start with the AWS Certified AI Practitioner or a degree program and come back to this one.

What Is Changing From MLA-C01 to MLA-C02?

This is the part to read before you pay for anything. The exam is mid-switch.

AWS lists September 28, 2026 as the last day to take the current exam, MLA-C01, in English. The updated version, MLA-C02, is in beta, and the beta is offered in English only. AWS says the new version adds generative AI implementation (including Amazon Bedrock and retrieval-augmented generation architectures), agentic AI and foundation models.

Here is how the two compare on the figures AWS gives:

• MLA-C01: $150 USD, 65 questions (multiple choice and multiple response), 130 minutes, passing score 720.
• MLA-C02 beta: $75 USD at beta pricing, 85 questions, 170 minutes.

What that means for you:

• If you’re studying now, study the current guide’s domains and also read what AWS says is new in MLA-C02. Generative AI is clearly headed into this exam.
• If you want the credential soon, open AWS’s certification page and see which version you can book today. We can’t tell you what’s bookable on a given day.
• If you hold the older version, the credential is valid for 3 years from when you earned it.

The passing score of 720 applies to MLA-C01.

How Much Does the Exam Cost?

MLA-C01 costs $150 USD. The MLA-C02 beta is listed at $75 USD under beta pricing. Beta pricing is a temporary offer, so it may change when the exam leaves beta.

Compare that with the other AWS AI exams. The AI Practitioner is $100, and the Generative AI Developer Professional is $300. The retired Machine Learning Specialty was $300 at the end.

The exam fee is the smaller part of the real cost. The larger cost is the experience AWS recommends, which takes months of hands-on work to build.

What Is the Exam Format?

For MLA-C01, AWS publishes this format:

• Questions: 65, multiple choice and multiple response.
• Time: 130 minutes.
• Delivery: Pearson VUE testing center or online proctored.
• Scoring: scaled from 100 to 1,000. The minimum passing score is 720.

The MLA-C02 beta has 85 questions and runs 170 minutes. It’s English only while in beta.

Multiple-response questions mean you pick more than one correct answer. They’re a common place to lose points, because a partial answer may not earn credit. Read each question for how many answers it wants.

What Does the Exam Cover?

The MLA-C01 exam guide lists four domains with these weights of scored content:

• Data Preparation for Machine Learning: 28%
• ML Model Development: 26%
• Deployment and Orchestration of ML Workflows: 22%
• ML Solution Monitoring, Maintenance, and Security: 24%

The weights are close, which is the point. This exam rewards someone who has seen the whole life of a model, not just the modeling step.

Data preparation is the largest domain. Expect questions about getting data into usable shape, cleaning it and choosing how to store and transform it. If you’ve done real ML work, you know this step eats most of the time.

Model development covers picking an approach, training, tuning and evaluating.

Deployment and orchestration is about getting a model into production and automating the workflow around it.

Monitoring, maintenance and security covers what happens after launch, including keeping the model accurate and protected.

These are the domains from the MLA-C01 guide. AWS says MLA-C02 adds generative AI and agentic AI topics, so expect the new guide’s mix to shift.

Who Can Take It, and What Experience Do You Need?

AWS says the ideal candidate has at least 1 year of experience in machine learning engineering or a related field and 1 year of hands-on experience with AWS services. It also notes that you don’t need prior ML experience to start preparing. No degree is required.

Read that as a recommendation, not a gate. Nobody checks your résumé before you book. But the exam questions assume you’ve seen these problems, and a candidate with no hands-on work will find the scenario questions hard.

How Do You Prepare?

Start with the exam guide, because it lists the tasks and skills under each domain. Then build the experience the guide assumes.

1. Read the guide and mark gaps. Be honest about which tasks you’ve never done.
2. Work the heaviest domain first. Data preparation carries 28% in the current guide.
3. Build one small project end to end on AWS. Prepare data, train a model, deploy it and watch it. One finished pipeline teaches more than ten videos.
4. Learn the deployment side. Many people prepare well on modeling and thin on orchestration and monitoring, which together make up nearly half the scored content.
5. Read what is new in MLA-C02. Look at generative AI and foundation model topics, including Amazon Bedrock and RAG.
6. Practice with exam-style questions. Focus on multiple-response items.

Give yourself more time than you’d give the foundational exam. The gap in difficulty is real.

How Does It Relate to an AI Degree?

They aren’t substitutes. A cert doesn’t replace a degree, and a degree doesn’t replace a cert.

A degree teaches the theory, statistics and programming behind machine learning, and it gives you college credit and a transcript. The associate exam checks that you can operate ML workflows on one cloud. It won’t give you the math background that research-style or senior roles ask for, and no hiring manager will read it as a degree.

A degree also won’t teach you AWS by default. Many programs teach the concepts and leave the platform to you. This exam fills that gap.

The pairings that make sense:

• Degree plus the exam. The strongest résumé for engineering roles, since it shows depth and platform skill.
• Experience plus the exam. Works for people already in software or data jobs.
• The exam alone. Fine for a promotion or a project, weak for a career change into ML.

Our ranking lists cover the degree side. Online machine learning degrees and online AI engineering degrees match this exam’s engineering focus most closely. If you want graduate depth, see online master’s in artificial intelligence. For the bigger question of whether a degree pays off, read Is an AI Degree Worth It?.

What Jobs Does It Help With?

The credential points at machine learning engineering work. See our pages on machine learning engineer and AI engineer jobs for duties, degree levels and pay. Related roles include data engineer and software engineer positions that now include model deployment.

No exam promises a job. What it does is give a recruiter a clear signal that you’ve worked on the full ML workflow on AWS. Employers that run on AWS may care. Employers that run on another cloud may not.

Pull up postings for the job you want and look for AWS in the requirements. That tells you whether the exam helps.

How Does It Compare With the Other AWS AI Exams?

AWS offers a ladder, and this exam sits in the middle.

• AWS Certified AI Practitioner: foundational, $100, 65 questions, 90 minutes, no prerequisite. See the AI Practitioner page.
• AWS Certified Machine Learning Engineer Associate: this exam, $150 for MLA-C01, 130 minutes, about a year of experience recommended.
• AWS Certified Generative AI Developer Professional: $300, 75 questions, 180 minutes, with 2 or more years of production experience recommended. Our Generative AI Developer page has the details.

Pick by what you’ve done, not by what sounds impressive. A person with no hands-on experience who sits a professional exam is paying $300 to learn what they’re missing. The associate exam is the right level for people who have shipped a model or two.

What Does a Realistic Study Plan Look Like?

The plan depends on where you start.

You already work in software or data. You have the AWS familiarity and probably some of the pipeline skills. Spend your time on the gaps, which are usually model deployment and monitoring. Build one small project on AWS and read the exam guide for terms you haven’t met.

You’re a student in an AI degree. You likely know the modeling but may not have touched deployment on a cloud platform. Use a course project or an internship to get hands-on time, then study the exam guide.

You’re newer to the field. Don’t force this exam yet. The AI Practitioner is a gentler start, and a degree or a year of project work will make the associate exam far more manageable. Rushing it can mean paying for a retake.

In every case, allow time for the MLA-C02 changes. If you study from older material only, you may miss the generative AI topics.

What Mistakes Do Candidates Make?

Treating it as a modeling exam. It’s an engineering exam. Data preparation, deployment and monitoring carry a large share of the scored content, so a strong modeler can still be caught short.

Studying a version you can’t book. With the switch underway, confirm which exam you’re preparing for before you buy a prep course.

Ignoring security. It’s part of the fourth domain, along with monitoring and maintenance.

Skipping hands-on time. Reading about a pipeline is not the same as building one. The scenario questions reward people who have seen things fail.

What About the Machine Learning Specialty?

The older AWS Certified Machine Learning Specialty was retired, and the last day to take it was March 31, 2026. AWS points to this associate certification as its option. If you’re researching the old credential, our page on what to take instead of the ML Specialty covers it.

How Long Does It Last?

The certification is valid for 3 years. You recertify by passing the latest version of the exam. Earning this credential also automatically recertifies the AWS Certified AI Practitioner.

That’s useful if you plan to climb the ladder. The foundational credential and the associate credential keep each other current.