The AWS Certified Machine Learning Engineer Associate (MLA-C02) is worth it in 2026 if you build, deploy or operate ML or generative AI workloads on AWS and want a current credential that proves it. It is not worth it as a first cloud certification or a stand-in for experience. The C02 refresh decides which side you are on.
Last updated October 2026.
Money first, then time. The standard AWS associate exam fee is $150 USD, with the MLA-C02 beta offered at a reduced price, and every retake is paid in full. The exam is 65 questions in 130 minutes (50 of them scored), a passing score of 720 on the 100 to 1,000 scale, and compensatory scoring, so one weak domain can be carried by the others. The credential is valid for three years.
Time is the bigger line item. The MLA-C02 study plan puts the typical candidate at 10 to 12 weeks of part-time study, roughly 80 to 90 hours. Engineers who already ship on both Amazon SageMaker AI and Amazon Bedrock often need 4 to 6 weeks; someone new to both toolchains should budget 14 to 16.
One caveat: beta results are reported after the beta closes, a delay the cost and format post explains.
MLA-C01 was retired in English on 28 September 2026 and MLA-C02 delivery began on 29 September 2026. The format barely moved: the same 65 questions, 130 minutes and 720 to pass, with the domain weights now 28/24/24/24. The scope moved a lot. AWS did not add a generative AI domain; it added foundation models, retrieval-augmented generation, agents and guardrails inside all four existing domains, so there is no GenAI corner to skip. The MLA-C01 vs MLA-C02 comparison lists every added and removed skill.
The exam guide now describes its target candidate as someone with at least one year using Amazon SageMaker AI, Amazon Bedrock and other AWS services for ML engineering, plus a year in a related role such as data engineer, backend developer or data scientist. In practice that means the certification is for people who do both halves of the job: train and deploy traditional models, and stand up Bedrock-based RAG and agent workloads. If your work is only one of those, the other half is where your study hours go, and that decides whether the time is worth it.
An engineer who puts ML or generative AI into production on AWS and wants a credential that reflects the 2026 version of that job; a data engineer on AWS moving toward ML platform work; a data scientist who wants to prove deployment, monitoring and security skills, not just modelling; or anyone whose employer values or reimburses AWS certifications.
New to AWS or to ML, where a foundational exam is the better first step; a Databricks or other-platform practitioner with no AWS roadmap; a senior engineer whose shipped systems already speak louder than a badge; or someone hoping the certificate alone lands an ML engineering role without projects behind it.
The honest answer is situational. These are the signals that tip it one way or the other:
| Your situation | Verdict | Why |
|---|---|---|
| You ship models on SageMaker AI and build on Bedrock | Yes | You are the target candidate. Four to six weeks of review and the exam confirms what you already do. |
| You are a data engineer on AWS heading toward ML | Yes, with a 10 to 12 week plan | Domain 1 (28%) is close to your day job; the FM development and ML monitoring topics are the runway. |
| You build or evaluate models but never deploy them | Maybe | Domains 3 and 4 are 48% of the exam. Worth it if you want to grow into deployment; a long road if you do not. |
| You are new to AWS and to ML | Not yet | Start with the AI Practitioner (AIF-C01) or another foundational exam. MLA-C02 assumes two toolchains you have not met. |
| Your platform is Databricks, Azure or GCP | Skip | Take that platform's credential. The AWS service names are most of the study load and will not transfer. |
| You need the credential on a fixed date during the beta | Weigh the delay | Beta results arrive after the beta closes. If the date is hard, plan around it or wait for the standard release. |
Format, weights and target-candidate description from the official MLA-C02 exam guide; timelines are Certified's estimates from the study-plan post.
The realistic take: if ML or generative AI on AWS is your work or your next role, the $150 and ten weeks are an easy yes. If it is neither, the same hours go further on the platform you actually use.
Neither exam is a prerequisite for the other. The AWS Certified AI Practitioner (AIF-C01) is a foundational exam that tests AI, ML and generative AI concepts in 90 minutes for $100; MLA-C02 tests whether you can build and operate them. If you already write Python and use SageMaker AI or Bedrock at work, AIF-C01 would mostly confirm what you know, so go straight to the associate exam. If AI on AWS is new to you, AIF-C01 builds the vocabulary that now runs through every MLA-C02 domain. The AI Practitioner vs ML Engineer comparison walks through the choice.
Yes, if you work on AWS and want to move toward ML platform work. Domain 1, Data Preparation for ML and AI, is 28% of the exam and close to a data engineer's day job, while the foundation model development and ML monitoring topics are where the new study time goes. Plan on the typical 10 to 12 weeks.
Not as a requirement, but it changes the effort. The exam guide describes a target candidate with a year on both Amazon SageMaker AI and Amazon Bedrock, and the C02 additions (foundation models, RAG, agents, guardrails) sit in all four domains. With Bedrock experience the exam confirms your skills; without it, Bedrock is most of the study load.
The standard AWS associate fee is $150 USD, and AWS prices beta exams lower. Each retake is paid in full, and the certification is valid for three years. Confirm the current price and language on the registration page, since beta and standard terms differ.
Only if AI on AWS is new to you. AWS has no prerequisites for either exam, so engineers who already use SageMaker AI or Bedrock can sit MLA-C02 directly. AIF-C01 is a $100, 90-minute foundational ramp that builds vocabulary, not a gate.
Yes. AWS certifications, including the Machine Learning Engineer Associate, are valid for three years, after which you recertify against the version of the exam that is current at the time. From 29 September 2026 the English exam offered is MLA-C02, which replaced MLA-C01.
The quickest way to know whether MLA-C02 is worth it for you is to work through the material. Certified's course maps all four domains to 76 short chapters with a practice exam per unit and three full-length mocks. Unit 1 is free to start.
Open the MLA-C02 course →