MLOps Engineering on AWS

This course introduces Machine Learning Operations (MLOps) and extends DevOps principles to the full lifecycle of machine learning models, from data preparation and training through deployment, monitoring, and continuous improvement. Participants learn how to operationalize ML models at scale by integrating data, code, and models using automation, collaboration, and AWS-native tools. The course emphasizes real-world challenges such as cross-team handoffs, model deployment strategies, monitoring for data drift and bias, and taking corrective action when production models deviate from expected performance. By the end of the course, learners will be able to design and implement automated, production-ready ML workflows using Amazon SageMaker and related services.

• Explain the principles and goals of machine learning operations (MLOps)
• Distinguish between DevOps and MLOps workflows and challenges
• Describe the end-to-end machine learning lifecycle from development to production
• Design automated ML workflows for building, training, testing, and deploying models
• Use Amazon SageMaker features to support MLOps automation and governance
• Implement CI/CD-style pipelines for ML model retraining and redeployment
• Package and deploy models using appropriate inference and scaling strategies
• Select deployment options for real-time, batch, and edge inference
• Monitor models for data drift, bias, performance, resource usage, and latency
• Integrate human-in-the-loop reviews into production ML systems
• Apply best practices for communication, collaboration, and operational ownership in MLOps

• ML data platform engineers
• DevOps engineers supporting machine learning workloads
• Developers and operations staff responsible for deploying and maintaining ML models
• Technical professionals involved in ML pipeline automation and production ML systems

• AWS Technical Essentials
• DevOps Engineering on AWS
• Practical Data Science with Amazon SageMaker
• Familiarity with basic machine learning concepts and cloud-based workflows