CI/CD for Multi-Model Endpoints in AWS

A simple, flexible alternative for sustainable ML solutions

Automating the retraining and deployment of production machine learning solutions is a pivotal step to ensure models account for covariate shift while limiting error-prone and unnecessary human effort.

For models deployed using the AWS stack and particularly SageMaker, AWS offers a standard CI/CD solution using SageMaker Pipelines to automate retraining/deployment, and the SageMaker Model Registry to track lineage of a model.

While the standard …

aws ci-cd-pipeline machine learning ml-pipeline sagemaker