Posted On: Aug 4, 2020

You can now include Amazon SageMaker Processing in your machine learning workflows created using AWS Step Functions, allowing you to build data processing and compute steps into your machine learning workflows without leaving the SageMaker service.

AWS Step Functions allows you to build resilient workflows using AWS services such as Amazon DynamoDB, AWS Lambda, and Amazon SageMaker. Amazon SageMaker helps developers and data scientists build, train, and deploy machine learning (ML) models quickly. Now, with the integration of SageMaker Processing into Step Functions, you can orchestrate end to end machine learning workflows that include data pre-processing, post-processing, feature engineering, data validation, and model evaluation on Amazon SageMaker.

You can now use the Step Functions Data Science SDK with Amazon SageMaker Processing to create and visualize end-to-end machine learning workflows. Workflows can be built in Python and visualized within Jupyter Notebooks. Data scientists can build and iterate on their machine learning pipelines and then write out a CloudFormation template that can be used by engineering teams to take the workflow into production, supporting the MLOps use-case.

AWS Step Functions support for Amazon SageMaker Processing is available in all regions where Step Functions and SageMaker Processing are available. View the AWS Regions table to learn more. To learn more about Amazon SageMaker Processing, view our blog post, read the SageMaker Developer Guide, read the Step Functions Developer Guide, and try a Step Functions sample project.