Launch Amazon SageMaker Autopilot experiments from Amazon SageMaker Pipelines to easily automate MLOps workflows

Amazon SageMaker Autopilot, a low-code machine learning (ML) service which automatically builds, trains, and tunes the best ML models based on your data, is now integrated with Amazon SageMaker Pipelines, the first purpose-built continuous integration and continuous delivery (CI/CD) service for ML. This enables the automation of an end-to-end flow of building ML models using SageMaker Autopilot and integrating models into subsequent CI/CD steps.

Starting today, you can add an automated training step (AutoMLStep) in SageMaker Pipelines and invoke a SageMaker Autopilot experiment with Ensemble training mode. As an example, let’s consider building a training and evaluation ML workflow for a fraud detection use case with SageMaker Pipelines. You can now launch a SageMaker Autopilot experiment using the AutoML Step which will automatically run multiple trials to find the best model on a given input dataset. After the model package for the best model is created using the CreateModel step, its performance can be evaluated on test data using the Transform step within SageMaker Pipelines. Eventually, the model can be registered into the SageMaker Model Registry using the RegisterModel step.

Native support for SageMaker Autopilot as a step within SageMaker Pipelines is now available in all regions where SageMaker Pipelines is available except Amazon Web Services China (Beijing) Region and Amazon Web Services China (Ningxia) Region. To learn more about SageMaker Pipelines and SageMaker Autopilot, visit the SageMaker Pipelines product page and SageMaker Autopilot product page.