Creating CI/CD Pipeline for Machine Learning (ML)

6 hours of instruction

This practical, hands-on course recaps and ties together all stages of ML cycle in production into an automated CI/CD pipeline.

OBJECTIVES

1. By the end of this course, participants will be able to tie all stages of ML and create automated pipelines on AWS

PREREQUISITES

MLOPs CI/CD Theory

SYLLABUS & TOPICS COVERED

1. Automating model deployment
   • Summarize the characteristics of ML pipeline automation
   • Describe model deployment in SageMaker
   • Configure model deployment in SageMaker
2. Model deployment file structure
   • Explain the Model Deploy build files and their uses
   • Identify the Model Deploy test files and their uses
3. Developing a code pipeline
   • Explain the Source stage and analyze the benefits of CodeCommit
   • Describe the Build stage and evaluate the importance of CodeBuild
4. Executing a code pipeline
   • Examine the benefits of using CloudFormation templates
   • Implement each step in the DeployStaging stage
   • Execute the DeployProd stage
5. Connecting SageMaker models to APIs
   • Test API endpoints using the SageMaker console
   • Test API endpoints using SageMaker SDK
6. Serverless computing with AWS Lambda
   • Explain AWS Lambda and outline how it works
   • Create a Lambda function
   • Configure a Lambda function

SOFTWARE REQUIREMENTS

API Gateway, AWS Sagemaker, Access to AWS accounts, CodeBuild, CodePipeline, Lambda, S3