Exam Domains¶
The AWS Certified Machine Learning Engineer Associate exam covers four content domains.
Domain 1: Data Preparation for Machine Learning (28%)¶
This domain focuses on ingesting, transforming, validating, and preparing data for ML modeling.
Task Statements¶
- Ingest and store data for ML workloads
- Transform data and perform feature engineering
- Ensure data integrity and prepare data for modeling
Key Services¶
- Amazon S3, AWS Glue, AWS Glue DataBrew
- Amazon Kinesis, Amazon Data Firehose
- AWS Lake Formation, Amazon Athena
- SageMaker Data Wrangler, Feature Store
Domain 2: ML Model Development (26%)¶
This domain covers model selection, training, tuning, evaluation, and versioning.
Task Statements¶
- Choose modeling approaches based on business objectives
- Train and refine ML models
- Analyze model performance and versions
Key Services¶
- Amazon SageMaker (Training, Built-in Algorithms)
- SageMaker Experiments, Debugger, Clarify
- SageMaker Model Registry
- Amazon Bedrock
Domain 3: Deployment and Orchestration of ML Workflows (22%)¶
This domain focuses on deploying models and setting up CI/CD pipelines.
Task Statements¶
- Select deployment infrastructure and configure endpoints
- Create and script infrastructure for ML models
- Use CI/CD pipelines to automate workflows
Key Services¶
- SageMaker Endpoints (Real-time, Serverless, Async)
- SageMaker Pipelines, AWS Step Functions
- Amazon ECR, ECS, EKS
- AWS CodePipeline, CodeBuild
Domain 4: ML Solution Monitoring, Maintenance, and Security (24%)¶
This domain covers monitoring, cost optimization, and security best practices.
Task Statements¶
- Monitor model performance and data quality
- Monitor and optimize infrastructure and costs
- Secure AWS resources for ML
Key Services¶
- SageMaker Model Monitor
- Amazon CloudWatch, CloudTrail
- AWS Cost Explorer, Budgets
- IAM, KMS, Secrets Manager, Macie