AWS AI Certifications: AIF-C01, MLA-C02 and AIP-C01

Amber enrol buttons are Udemy affiliate links; we earn a commission if you enrol through them. How we're funded.

Where we would start, among the ones that pay us

AWS's AI certifications changed shape in 2026. It retired the Machine Learning – Specialty, updated the Machine Learning Engineer – Associate as MLA-C02 with generative and agentic AI added, and now has a business-track exam in beta. This guide sets out what is live, taken from AWS's own certification pages and exam guides as we read them on 24 September 2026: what each exam covers, who AWS writes it for, which to take first, and where AWS's free preparation ends and paid preparation is worth adding.

What AI certifications does AWS offer?

AWS's technical AI path has three certifications, one at each level: AI Practitioner at Foundational, Machine Learning Engineer at Associate and Generative AI Developer at Professional. All three are taken through Pearson VUE, at a test centre or as an online proctored exam, and AWS asks you to recertify every three years.

The table below shows the fee and format of the three exams this page discusses, as AWS publishes them. Every other exam we track is compared on our AI certification exams page.

The Level column uses this site’s Foundational, Intermediate and Advanced scale; AWS’s own categories for the same three exams are Foundational, Associate and Professional.

In AWS's own words, the three test different jobs:

• AWS Certified AI Practitioner “validates foundational knowledge of artificial intelligence (AI), machine learning (ML), and generative AI concepts and use cases for AWS Services.”
• AWS Certified Machine Learning Engineer – Associate “validates technical ability in implementing ML workloads in production and operationalizing them.”
• AWS Certified Generative AI Developer – Professional “showcases advanced technical expertise in building and deploying production-ready AI solutions using AWS Services like Bedrock.”

A fourth AI certification sits outside that path. AWS Certified AI Business Strategist, currently a beta exam, “tests your ability to evaluate AI investments, build business cases, design governance, and scale adoption across an organization”, and AWS says it “does not assess AWS services knowledge”. It is written for product and program managers, sales and business development staff, consultants, business analysts and marketers, with “no coding or AWS implementation experience required”. It is not in the table above because its exam page and its exam guide give different exam lengths, so check both before you book.

AWS also runs microcredentials, which its exam pages list beside these certifications: timed, hands-on challenges in a live AWS environment, such as AWS Agentic AI Demonstrated and AWS MLOps Demonstrated. AWS describes them as complementing a certification by “demonstrating your practical implementation abilities”; they do not replace one.

Which AWS AI certification should you take first?

It depends less on ambition than on what you already do at work. AWS writes each exam for a stated level of experience, and sitting one before you have that experience risks paying for a retake.

The table below pairs five starting points with the AWS AI exam to take first and the one to aim for next, and gives the reason from AWS's own exam pages and guides.

For the wider choice by job, see our guides for cloud engineers, data engineers, software engineers and product managers. If you want no code at all, our no-coding AI certifications page sets AI Practitioner beside the other vendors' no-code exams.

AWS Certified AI Practitioner (AIF-C01)

AWS calls AI Practitioner “a good starting point for individuals exploring AI/ML on AWS, whether adding AI skills to an existing cloud career or beginning to work with AI/ML technologies.” Its exam guide describes a candidate with “up to 6 months of exposure to AI/ML technologies on AWS” and lists “Developing or coding AI/ML models or algorithms” among the tasks that are out of scope, which is why it is the no-coding exam on AWS's technical AI path. The five domains run from the fundamentals of AI, machine learning and generative AI to applications of foundation models, responsible AI, and security, compliance and governance for AI solutions.

After it, AWS recommends the Data Engineer – Associate or the Machine Learning Engineer – Associate for anyone heading into data, AI or machine learning, and earning the Machine Learning Engineer – Associate automatically recertifies AI Practitioner. Our AWS AI Practitioner review weighs whether it is worth the fee, the AIF-C01 study guide has a week-by-week plan, and AWS AI Practitioner vs Microsoft's fundamentals exam helps if you are choosing between the two clouds' entry points.

Machine Learning Engineer – Associate: MLA-C01 becomes MLA-C02

This exam is changing version. AWS set 28 September 2026 as the last day to take MLA-C01 in English. The updated exam, MLA-C02, opened for beta registration on 1 September 2026, and beta delivery begins on 29 September 2026, in English only. MLA-C01 stays available in Japanese, Korean and Simplified Chinese until MLA-C02 is generally available.

AWS says the domain structure “remains the same. No new domains were added.” What MLA-C02 adds is generative AI implementation with Amazon Bedrock and retrieval-augmented generation, agentic AI, the selection, fine-tuning and operation of foundation models and LLMs, and responsible AI practice across both traditional ML and generative AI. The experience it expects has moved with it: MLA-C02's guide asks for “at least 1 year of experience using Amazon SageMaker AI, Amazon Bedrock, and other AWS services for ML engineering”, a year in a related role such as backend software developer, DevOps developer, data engineer or data scientist, and experience “with both traditional ML and generative AI”.

The beta is longer than MLA-C01 and has more questions, as the table above shows, because beta exams “contain additional items used for statistical evaluation”; AWS says those items do not affect your score and that beta results are typically available within five business days. A beta can be taken only once: AWS’s exam policies say a candidate who fails must wait until the standard version is generally available, and the MLA-C02 page gives no date for that yet, so book the beta only when you are ready. The page’s exam-code row reads ME1-C02 for the same exam, so expect either string when you book. Whichever version you pass, AWS's page is explicit that “your certification remains active through its full validity period regardless of when you earn it.”

If you are starting now, prepare for MLA-C02: material built only for MLA-C01 misses the generative and agentic AI content. Our ML Engineer Associate study guide has the preparation plan, Databricks vs AWS vs Azure ML compares SageMaker with the other two platforms employers use for the same work, and our MLOps certifications guide covers the operations side the exam leans on.

Generative AI Developer – Professional (AIP-C01)

AWS describes this one as “perfect for developers with 2+ years of cloud experience”. Its exam guide is more specific: 2 or more years building production-grade applications on AWS or with open-source technologies, general AI/ML or data engineering experience, and 1 year of hands-on experience implementing generative AI. It tests building with vector stores, retrieval-augmented generation and knowledge bases, integrating foundation models into applications, prompt engineering, agentic AI, cost and performance optimisation, security and governance, and troubleshooting.

Model development and training are out of scope. This is an exam for people who build applications on top of foundation models, not for people who train them, and that is the clearest line between it and the ML Engineer exam. No certification is required first, though AWS says candidates “could benefit from” AI Practitioner, Solutions Architect – Associate, Machine Learning Engineer – Associate or Data Engineer – Associate beforehand. For the wider field, see our generative AI certifications and agentic AI certifications guides and how to become a generative AI engineer.

What happened to the Machine Learning – Specialty?

AWS retired it. Its page says: “The last day to take this exam is March 31, 2026. Certification holders will still have an active certification for 3 years from the date it was earned.” If you hold it, you need do nothing, and it belongs on your CV with its expiry date until then.

AWS names no direct successor. On the retired exam's own page, its answer to “What does AWS Training & Certification offer for Machine Learning?” is the Machine Learning Engineer – Associate, whose MLA-C02 version now covers generative AI as well as traditional ML. The Specialty was written for people with two or more years of running ML workloads on AWS; the AWS AI exam now pitched at that level of experience is the Generative AI Developer – Professional, though it tests building applications on foundation models rather than training models. One AWS page has not caught up: the ML Engineer exam page still suggests the Specialty as the next step after it, and that advice no longer applies.

How to prepare: AWS's free prep, and where Udemy fits

Start with what AWS gives away. Every exam page links an Exam Prep Plan on AWS Skill Builder, AWS's online learning centre, and AWS's certification-prep page lists what is free there: Official Practice Question Sets, 20-question sets that “demonstrate the style of our certification exams”, and Exam Prep courses, short digital courses that “explore an exam's topic areas and review sample certification questions for each domain.” The full-length Official Practice Exams, labs and extra practice questions need a paid Skill Builder subscription. For AI Practitioner, AWS also includes “free AI foundational training” in the Exam Prep Plan.

Paid preparation earns its place in two spots, and we recommend one Udemy product for each:

• For AIF-C01: AWS Certified AI Practitioner AIF-C01 Practice Exams, by Stephane Maarek and Abhishek Singh. It is a question bank, not a course: four full practice exams, 260 questions in all, and no video. Use it after you have studied, to find the gaps a single 20-question set is too small to show. Our review of these practice exams sets out what they do and do not do.
• For MLA-C02: AWS Certified Machine Learning Engineer Associate: Hands On!, from Sundog Education's Frank Kane and Stephane Maarek. It is a taught course of about 25 hours, running from data ingestion and SageMaker's built-in algorithms through training and tuning to building generative AI applications with Bedrock and MLOps. Its Udemy page says it has been fully updated for MLA-C02, and it includes two practice tests.

For AIP-C01 we do not yet recommend a paid course. The Hands On! course's Bedrock section touches part of that syllabus, but it was built for the Associate exam, so AWS's own Exam Prep Plan is the place to start. Both Udemy products are one-off purchases whose price moves with Udemy's sales, so check it on the day. Each ends in Udemy's certificate of completion; the AWS certification comes only from passing AWS's proctored exam. Our guide to using practice exams well explains how to read your scores.

Retakes, renewal and your second exam

If you fail a standard exam, AWS makes you wait 14 calendar days before a retake. There is no limit on attempts, but each one costs the full exam fee. A beta is the exception: AWS’s exam policies allow one attempt at a beta, and after a fail you must wait until the standard version is generally available. That covers the MLA-C02 beta and the AI Business Strategist beta. A certification lasts three years, and you recertify by passing the latest version of its exam; for AI Practitioner, passing the Machine Learning Engineer – Associate does it automatically. Once you hold an AWS certification, AWS's certification FAQ says you get access to “a 50% discount on additional exams when you recertify or upgrade”, as a voucher in the Benefits section of your AWS Certification Account, so the first certification you earn makes the next exam cheaper.

AWS or another cloud?

An AWS certification counts for most where employers run AWS. If you have not settled on a cloud, our comparison of AWS vs Azure vs Google Cloud AI certifications sets the three vendors' paths side by side, AI certification exams ranked by difficulty places AI Practitioner and the ML Engineer exam among the rest, and our AI certification exams page lists the fee and format of every vendor exam we track.

Ready to start?

Frequently asked questions