AWS Machine Learning Engineer Associate (MLA-C01)

Ingesting, transforming, cleaning, and preparing data and features for ML on AWS.

• Ingesting data from batch and streaming sources (Amazon S3, Kinesis, Glue)
• Data formats and storage for ML (Parquet, RecordIO, S3 partitioning)
• Data cleaning: handling missing values, outliers, duplicates
• Feature engineering, encoding, scaling, and transformation
• Amazon SageMaker Data Wrangler and Feature Store
• AWS Glue and Glue DataBrew for ETL
• Splitting data into train/validation/test; handling class imbalance
• Detecting and mitigating data bias (SageMaker Clarify)
• Data labeling with Amazon SageMaker Ground Truth

Free Data Preparation for Machine Learning practice questions →Data Preparation for Machine Learning study guide →

Choosing algorithms, training, tuning, and evaluating machine learning models.

• Selecting appropriate algorithms and built-in SageMaker algorithms
• Training models with SageMaker (training jobs, script mode, built-in algorithms)
• Hyperparameter tuning (SageMaker Automatic Model Tuning)
• Evaluation metrics: accuracy, precision, recall, F1, AUC, RMSE, MAE
• Confusion matrix and interpreting classification results
• Overfitting/underfitting, regularization, cross-validation
• Using SageMaker JumpStart and pretrained/foundation models
• Managing experiments and model versions (SageMaker Experiments, Model Registry)

Free ML Model Development practice questions →ML Model Development study guide →

Deploying models and building automated, scalable ML pipelines on AWS.

• SageMaker inference options: real-time endpoints, serverless, asynchronous, batch transform
• Choosing deployment based on latency, throughput, and cost
• Auto scaling inference endpoints
• Deployment strategies: blue/green, canary, shadow testing
• Building ML pipelines (Amazon SageMaker Pipelines)
• Infrastructure as code for ML (AWS CloudFormation, CDK)
• CI/CD for ML (SageMaker Projects, code repositories, automation)
• Containerizing models (bring-your-own-container, Amazon ECR)

Free Deployment and Orchestration of ML Workflows practice questions →Deployment and Orchestration of ML Workflows study guide →

Monitoring model quality and data drift, maintaining models, and securing ML systems.

• Monitoring models in production (Amazon SageMaker Model Monitor)
• Detecting data drift and model/quality drift; triggering retraining
• Monitoring infrastructure and endpoints (Amazon CloudWatch)
• Logging and auditing (AWS CloudTrail)
• Cost monitoring and optimization for ML workloads
• Securing ML systems: IAM least privilege, encryption, VPC isolation
• Data protection and privacy for ML pipelines
• Explainability and bias monitoring (SageMaker Clarify)

Free ML Solution Monitoring, Maintenance, and Security practice questions →ML Solution Monitoring, Maintenance, and Security study guide →