SageMaker Pipelines for End-to-End Model Deployment

This project demonstrates the use of Amazon SageMaker Pipelines to run end-to-end workflows for model deployment. The pipeline includes steps for preprocessing, training, and creating and registering a model.

Features

• Preprocessing: Data preparation and cleaning steps.
• Training: Model training with specified algorithms and parameters. The trained model artifacts are saved in an S3 bucket for later use.
• Model Creation and Registration: Steps to create and register the model in the model registry. Users can manually change the approval status of the model.

Workflow Overview

When the model's approval status is changed to "approved," an event is triggered via AWS EventBridge. This event invokes a Lambda function that deploys the model using the saved artifacts.

Architecture

Model Registration

• Version Control: Track different versions of your models.
• Approval Workflow: Manage model approvals before deployment.

A user has the option to analyze the performance of a version of the model and then approve it for deployment.

Event Handling

Utilizes AWS EventBridge to respond to model package state changes (Pending --> Approved).

Lambda Functions

Leverage AWS Lambda for deployment of the approved model.