Deploy Amazon SageMaker pipelines using AWS Controllers for Kubernetes

This article provides a solution for deploying Amazon SageMaker pipelines using AWS Controllers for Kubernetes (ACK). It explores how DevOps engineers can use Kubernetes to manage the entire machine learning lifecycle, including training and inference, using the same toolkit.

Specifically, the article covers:

• Overview of the solution architecture
• Prerequisites for following along
• Installing the SageMaker ACK service controller
• Generating a pipeline JSON definition
• Creating and submitting a pipeline YAML specification
• Submitting the pipeline to SageMaker
• Creating and submitting a pipeline execution YAML specification
• Reviewing and troubleshooting the pipeline run
• Cleanup steps