Practical Data Science with Amazon SageMaker

Syllabus

Module 1: Introduction to Amazon SageMaker

• Overview of SageMaker components
• SageMaker Studio and notebook environments
• Key benefits and use cases

Module 2: Preparing Data for Machine Learning

• Data ingestion using Amazon S3
• Data wrangling and cleaning
• Feature engineering using SageMaker Data Wrangler

Module 3: Model Development and Training

• Built-in algorithms vs. custom models
• Using SageMaker prebuilt containers
• Training jobs and hyperparameter tuning
• Running distributed training at scale

Module 4: Model Evaluation and Metrics

• Evaluating model performance
• Cross-validation and confusion matrix
• Logging and tracking experiments

Module 5: Model Deployment and Inference

• Real-time vs. batch inference
• Deploying models as endpoints
• Model A/B testing and versioning
• Cost optimization and autoscaling

Module 6: Automation and Pipelines

• Introduction to SageMaker Pipelines
• CI/CD for ML workflows
• Integrating with AWS Step Functions and Lambda

Module 7: Monitoring and Governance

• Model monitor and data drift detection
• Logging with Amazon CloudWatch
• Auditability and security in SageMaker

Module 8: End-to-End Project

• Capstone project using real-world dataset
• Full lifecycle: data prep, model training, deployment, and monitoring
• Peer review and instructor feedback