MLA-C02 exam domains and weighting

Exam format at a glance

All figures from the official AWS Certified Machine Learning Engineer Associate (MLA-C02) exam guide. The 15 unscored questions are not identified during the exam and do not affect your result.

The four content domains

The weights below are the share of scored content, so they describe how the 50 graded questions are distributed. Domain 1 leads at 28% and the other three are level at 24% each, a very flat spread. The practical read: the data domain is the single biggest, but nothing else is small enough to skim.

Domain 1: Data Preparation for ML and AI (28%)

The largest domain, and it is data engineering more than data science. It covers ingesting and storing data (S3, AWS Glue, Amazon EMR, streaming with Amazon Data Firehose and Amazon Managed Service for Apache Flink), transformation and feature engineering, and the AI-specific additions that define C02: embeddings, vector databases, chunking documents for retrieval-augmented generation (RAG), and SageMaker Feature Store.

Domain 2: ML and Foundation Model (FM) Development (24%)

Choosing a modeling approach, training and refining, and analyzing performance, now spanning both traditional models and foundation models. Expect built-in SageMaker algorithms and pre-trained models alongside FM selection, fine-tuning, continued pre-training, prompt engineering, and evaluating FM outputs, plus hyperparameter tuning and reading classification and regression metrics.

Domain 3: Deployment and Orchestration of ML and AI Workflows (24%)

Turning a trained model into a running service. It covers selecting deployment infrastructure for a set of requirements (real-time endpoints, serverless, asynchronous, and batch inference), sizing and autoscaling the compute, provisioning it as infrastructure as code, and wiring CI/CD and orchestration with SageMaker Pipelines and related AWS services.

Domain 4: Operating, Monitoring, and Securing ML and AI Solutions (24%)

Keeping a deployed solution healthy, affordable, and locked down. It covers monitoring model and agent performance and drift, observability and cost and performance tuning, and securing ML and AI systems with least-privilege IAM, network isolation, auditing with CloudTrail and AWS Config, and Amazon Bedrock Guardrails. Domain 3 plus Domain 4 is 48% of the exam, which is why MLA-C02 reads as an engineering credential.

What the weights mean in actual questions

Percentages are easier to plan against as question counts. With 50 scored questions, the four domains work out to roughly:

Counts are the published weights applied to 50 scored questions, rounded. AWS does not guarantee an exact per-domain count on any individual exam form.

How to use the weights when you study

With only a 4-point spread between the top and bottom domain, prioritising purely by weight barely helps. Prioritise by gap instead, and use the weights to decide when a domain is well enough covered to move on.

• Coming from data engineering? Domain 1 is close to your day job. Domains 2 and 4 hold the runway: FM development and evaluation, plus ML-specific monitoring and security.
• Coming from classic ML or data science? Domain 2 is home ground but only a quarter of the exam. Domains 3 and 4, together 48%, decide the result, and the GenAI topics (Bedrock, RAG, guardrails) are likely new.
• Read the task statements. The exam guide lists task statements and in-scope services under each domain, and questions are written from them. Reading them straight through is the fastest blind-spot audit available, and it is free.
• Build one end-to-end solution. Training a model, deploying it to an endpoint, adding monitoring, and scripting the whole thing touches all four domains at once.

FAQ