AWS Certified Machine Learning Specialty Training

This course prepares you for the MLA-C01 certification exam, covering all official exam domains and their approximate weightings:

Domain 1 Data Preparation for Machine Learning (ML) (28%)

• Ingest and store data from AWS sources (S3, EFS, FSx, Kinesis, Kafka, Flink) in formats like Parquet, JSON, CSV, ORC, Avro, RecordIO
• Merge data from multiple sources using AWS Glue, Apache Spark, or SageMaker Data Wrangler
• Transform data and perform feature engineering (scaling, binning, log transform, encoding, tokenization)
• Validate and label data using SageMaker Ground Truth or Amazon Mechanical Turk
• Identify and mitigate bias (class imbalance, difference in proportions of labels) using SageMaker Clarify
• Ensure data integrity, encryption, classification, and compliance (PII/PHI, data residency) before modeling

Domain 2 ML Model Development (26%)

• Choose a modeling approach: compare ML algorithms, SageMaker built-in algorithms, AI services (Rekognition, Transcribe, Bedrock), and foundation models via SageMaker JumpStart
• Train models using SageMaker built-in algorithms, script mode with TensorFlow/PyTorch, and fine-tune pre-trained models
• Tune hyperparameters (random search, Bayesian optimization) via SageMaker Automatic Model Tuning
• Prevent overfitting/underfitting/catastrophic forgetting via regularization, ensembling, stacking, boosting
• Manage model versions and repeatability using SageMaker Model Registry
• Evaluate model performance using metrics (F1, precision, recall, RMSE, ROC/AUC, confusion matrix) and SageMaker Clarify/Model Debugger

Domain 3 Deployment and Orchestration of ML Workflows (22%)

• Select deployment infrastructure and endpoint types (real-time, serverless, asynchronous, batch) based on cost/latency/performance tradeoffs
• Choose deployment targets (SageMaker endpoints, ECS, EKS, Lambda) and containers (built-in vs. BYOC)
• Provision and script infrastructure as code (CloudFormation, AWS CDK) and configure SageMaker endpoint auto scaling
• Optimize edge deployments using SageMaker Neo
• Set up CI/CD pipelines with CodePipeline, CodeBuild, and CodeDeploy for ML workflow automation
• Apply deployment/rollback strategies (blue/green, canary, linear) and build automated retraining and testing mechanisms

Domain 4 ML Solution Monitoring, Maintenance, and Security (24%)

• Monitor model inference in production for drift, data quality, and anomalies using SageMaker Model Monitor and Clarify
• Conduct A/B testing and compare shadow vs. production variant performance
• Monitor and optimize infrastructure and cost using CloudWatch, X-Ray, CloudTrail, Cost Explorer, and Trusted Advisor
• Rightsize instances and select purchasing options (Spot, On-Demand, Reserved, SageMaker Savings Plans)
• Secure ML resources via IAM least-privilege policies, roles, and SageMaker Role Manager
• Build VPCs, subnets, and security groups to isolate ML systems and secure CI/CD pipelines