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

About the MLA-C01 exam

AWS Certified Machine Learning Engineer – Associate (MLA-C01) tests whether you can take a machine learning model from raw data to a monitored production endpoint on AWS. It is an engineering exam, not a data science one. You are rarely asked to derive an algorithm. Most questions ask which SageMaker AI feature, data service or deployment pattern meets a requirement with the least effort, cost or risk.

The last day to sit MLA-C01 in English was September 28, 2026. Its successor, MLA-C02, has been delivered as a beta since September 29, 2026, and keeps the same certification name. MLA-C01 stays available in Japanese, Korean and Simplified Chinese until MLA-C02 reaches general availability, which AWS describes only as "early 2027".

Exam format

• Questions: 65, multiple choice and multiple response.
• Time: 130 minutes.
• Cost: 150 USD.
• Passing score: 720 on a scaled score from 100 to 1,000.

The four domains

• Data Preparation for Machine Learning (28%). Ingesting and storing data, transforming it with services such as AWS Glue and SageMaker Data Wrangler, feature engineering, and checking data quality and bias before training.
• ML Model Development (26%). Choosing an approach, training and tuning models in SageMaker AI, and evaluating them with the right metrics.
• Deployment and Orchestration of ML Workflows (22%). Picking an endpoint type, scaling it, and automating training and deployment with pipelines and CI/CD.
• ML Solution Monitoring, Maintenance, and Security (24%). Watching models and infrastructure in production, controlling cost, and securing data, models and endpoints.

What changes in MLA-C02

AWS kept the same four domains and the same task numbers. It renamed each domain to include AI as well as ML, and moved two points of weight from model development (now 24%) to deployment (now 24%). The new material is mostly generative AI engineering: foundation models on Amazon Bedrock, retrieval-augmented generation with vector stores and embeddings, agents, and evaluating, securing and controlling the cost of those systems. The beta has 85 questions in 170 minutes. The MLA-C01 to MLA-C02 guide covers every change.

How to use these questions now

This archive holds ten tests written for MLA-C01. They are kept as they are and no longer updated, but nearly everything in them is still on MLA-C02: data preparation, training and tuning, endpoints, pipelines, monitoring and security make up the same skeleton. Work through them for that core, read the explanation of every question you miss, and then add Bedrock, RAG and agents on top before you sit the new exam.

Some SageMaker AI features that appear in older study material are now closed to new customers, so check a feature's documentation page before relying on it in a new project, and read the current MLA-C02 exam guide on the AWS certification site before you book.

These questions are original, written from the public AWS documentation and the exam guide. They are not real exam questions, and this site is not affiliated with AWS.