MLOps Engineering on AWS
The MLOps Engineering on AWS course builds on and extends the DevOps practice prevalent in software development to building, training, and deploying ML models. The course emphasizes the importance of data, models, and code for successful ML implementations. It will demonstrate the use of tools, automation, processes and teamwork to address the challenges associated with handovers between data engineers, data scientists, software developers and operations.
COD: AW-MLOE
Categorie: AWS
Who should participate
- DevOps engineers
- ML engineers
- Developers/operators with responsibility for operationalizing ML models
Prerequisites
Completion of:
In this course you will learn to:
- Describe machine learning operations
- Understand the main differences between DevOps and MLOps
- Describe the machine learning workflow
- Discuss the importance of communications in MLOps
- Explain the end-to-end options for automating ML workflows
- List the main features of Amazon SageMaker for MLOps automation
- Create an automated ML process that builds, trains, tests and deploys models.
- Distribution operations
- Identify potential security threats in ML and explain basic mitigation approaches
- Describe why monitoring is important
- Detect drifts in the underlying input data
- Demonstrate how to monitor ML models for bias
- Explain how to monitor resource consumption and model latency.
Day 1
Module 1: Introduction
- Introduction to the course
Module 2: Introduction to MLOps
- Machine learning operations
- The goals of machine learning operations (MLOps)
- The path from DevOps to MLOps
- Machine learning
- Scope of
- An MLOps view of the machine learning workflow
- The comunication
- The value of MLOps: MLOps cases
Day 2
Module 3: MLOps development
- Introduction to building, training, and evaluating machine learning models
- Automate
- Apache Airflow
- Kubernetes integration for MLOps
- Amazon SageMaker per MLOps
- Demonstration: Amazon SageMaker
- Introduction to building, training, and evaluating machine learning models
- Demonstration: Lab overview
- Lab: Bringing your own algorithm into an MLOps pipeline
- Group activity: MLOps action plan notebook
- Lab: Coding and Serving Your ML Model with AWS CodeBuild
Module 4: Deploying MLOps
- Introduction to deployment operations
- Model packaging
- Inference
- Lab: Deployment of the model in production
- Production variants of SageMaker
- Distribution strategies
- Marginal distribution
- Distribution security
- Lab: Running A/B Tests
- Group activity: MLOps work plan
Day 3
Module 5: Model Monitoring and Operations
- The importance of monitoring
- Monitoring by design
- Lab: Tracking the ML Model
- The man in the cycle
- Amazon SageMaker Model Monitor
- Demo: Amazon SageMaker Model Monitor
- Solve problems
- Group activity: MLOps work plan
Module 6: Conclusion of the course
- Course review
- Group activity: MLOps Action Plan Workbook
- Conclusion
Duration – 3 days
Delivery – in Classroom, On Site, Remote
PC and SW requirements:
- Internet connection
- Web browser, Google Chrome
- Zoom
Language
Instructor: English
Workshop: English
Slides: English
