AWS Certified Machine Learning Engineer – Associate (MLA-C01 / MLA-C02)

By /

Vendor: Amazon Web Services · Last verified: 3 October 2026 · How we rate

Fee $150Level IntermediateAccess Open to anyoneDifficulty 3/5

1. The quick verdict

Worth it if you build, deploy or operate ML systems on AWS, mainly with SageMaker. It is now AWS’s main hands-on ML credential, since the Machine Learning – Specialty exam retired in March 2026.

The one thing to know before you register: a new version, MLA-C02, is in beta and adds foundation models and Amazon Bedrock, so check which version you are booking and study for that one.

2. Our ratings

CriterionScoreWhy
Difficulty (5 = hardest)●●●○○ 3/5Broad SageMaker and MLOps detail; manageable with real experience.
Cost burden, including fee, prep and retake risk (5 = most expensive)●●○○○ 2/5$150 (or $75 in beta) with meaningful prep time.
Career value today (5 = highest)●●●●○ 4/5The main AWS ML credential since the Specialty retired; well known to employers.
Skill transfer beyond this vendor (5 = most portable)●●●○○ 3/5ML lifecycle and MLOps thinking transfer; SageMaker specifics do not.
Hands-on experience required (5 = must have built real things)●●●●○ 4/5AWS recommends at least a year of hands-on SageMaker work.
Ease of access (5 = anyone can register)●●●●● 5/5Open to anyone.

3. At a glance

FieldDetails
LevelAssociate
AccessOpen to anyone
Exam feeMLA-C01: US$150. MLA-C02 beta: US$75.
Retake policyFull fee each attempt, 14-day wait, no attempt limit. Beta exams can be taken only once.
FormatPearson VUE, online proctored or test centre. MLA-C01: 65 questions (50 scored), 130 minutes, English, Japanese, Korean, Simplified Chinese. MLA-C02 beta: 85 questions, 170 minutes, English only. Multiple choice, multiple response, ordering and matching.
Passing score720 on a 100–1,000 scaled score
Coding neededReading code and config (SageMaker SDK, pipelines, IAM policies). No writing code.
Formal prerequisitesNone
Realistic prerequisitesAt least 1 year using SageMaker and related AWS ML services, plus a year in a role such as data scientist, data engineer, backend or DevOps developer. C02 adds Amazon Bedrock.
Validity & renewal3 years. Recertify by passing the latest version, passing Generative AI Developer – Professional, or a paid Skill Builder maintenance option. 50% discount voucher for recertification exams.
Estimated prep timeWorking with SageMaker now: 30–45 hours. ML practitioner new to AWS: 60–80 hours.
Official exam guideMLA-C01 exam guide (AWS)

Prep-time estimates are our own judgment, not an AWS figure.

4. What the exam tests

MLA-C01 (current version):

DomainWeightWhat it means in plain language
Data Preparation for ML28%Ingesting, cleaning, transforming and feature-engineering data for training.
ML Model Development26%Choosing algorithms, training, tuning and evaluating models.
ML Solution Monitoring, Maintenance, and Security24%Watching models in production for drift, cost and security issues.
Deployment and Orchestration of ML Workflows22%Deploying endpoints and automating pipelines with CI/CD and infrastructure as code.

MLA-C02 (beta) keeps the same four-part shape but widens each domain to cover foundation models: Data Preparation for ML and AI (28%), ML Model and Foundation Model Development (24%), Deployment and Orchestration of ML and AI Workflows (24%), and Operating, Monitoring, and Securing ML and AI Solutions (24%).

5. Who should take it, and who shouldn’t

Take it if you are:

  • An ML, MLOps or data engineer running ML workloads on AWS.
  • A data scientist who wants to prove production skills, not just modelling.
  • A backend or DevOps engineer moving into ML platform work.

Skip it or wait if you are:

  • Mainly building generative AI apps with foundation models. Generative AI Developer – Professional fits better.
  • New to AWS and ML. Start with AI Practitioner.
  • Not on AWS. The ML concepts transfer, but most questions are about SageMaker specifics.

6. What to expect after passing

  • What it signals: You can take an ML model from data to a monitored production endpoint on AWS.
  • What it doesn’t prove: Deep research-level ML, or generative AI app design (C01 barely covers it).
  • Where it’s asked for: ML engineer and MLOps job postings that use AWS, and AWS partner requirements. Since the Specialty retired, it is the main AWS ML credential employers will look for.
  • Natural next step: Generative AI Developer – Professional.

7. How to prepare

  • Official: The exam guide for your version, the AWS Skill Builder exam prep plan, and SageMaker documentation (training, endpoints, Pipelines, Model Monitor). For C02, add Amazon Bedrock.
  • Practice: Use scenario questions that ask for the right AWS service or configuration under cost, latency or security constraints. Make sure the set matches your exam version.

Hands-on:

  1. Train and deploy a model on SageMaker, then turn on Model Monitor (Model Development, Monitoring).
  2. Build a SageMaker Pipeline that retrains on new data, triggered from CI/CD (Deployment and Orchestration).
  3. Prepare a messy dataset with SageMaker Data Wrangler or Glue and record each transformation (Data Preparation).

8. Compare with

See all AWS guides → · Compare all AI certifications →

9. FAQ

MLA-C01 or MLA-C02: which should I take?

C02 is the future version and covers foundation models and Bedrock, but it is in beta: English only, results take weeks, and you can sit it once. If you need the credential soon, take C01; if you work with Bedrock and can wait, C02 is the better long-term choice.

Did this replace the Machine Learning – Specialty?

In practice, yes. The Specialty’s last exam date was 2026-03-31, and existing holders keep it until it expires. ML Engineer – Associate is now AWS’s main ML certification.

Do I need to write code?

No. You read SDK snippets, pipeline definitions and configurations and choose the right answer.

How do I keep it active?

It lasts 3 years. Passing Generative AI Developer – Professional also renews it.

Sources

Scroll to Top