Amazon SageMaker
Developer Guide

Custom Workflows via the API

When you have created your custom UI template (Step 2) and processing Lambda functions (Step 3), you should place the template in an Amazon S3 bucket with a file name format of: <FileName>.liquid.html.

Use the CreateLabelingJob action to configure your task. You'll use the location of a custom template (Step 2: Creating your custom labeling task template) stored in a <filename>.liquid.html file on S3 as the value for the UiTemplateS3Uri field in the UiConfig object within the HumanTaskConfig object.

For the AWS Lambda tasks described in Step 3: Processing with AWS Lambda, the post-annotation task's ARN will be used as the value for the AnnotationConsolidationLambdaArn field, and the pre-annotation task will be used as the value for the PreHumanTaskLambdaArn.