AWS Step Functions launches optimized integration for Amazon Bedrock
Today, AWS Step Functions announces an optimized integration for Amazon Bedrock, making it easier to integrate generative AI capabilities into Step Functions workflows, accelerating development of generative AI applications. This launch enables you to use Step Functions to orchestrate interactions with foundation models and add in steps for human intervention, without needing to write or maintain integration code. You can leverage Step Functions Workflow Studio to visually develop, inspect, and audit user defined generative AI workflows with ease.
Today, AWS Step Functions announces an optimized integration for Amazon Bedrock, making it easier to integrate generative AI capabilities into Step Functions workflows, accelerating development of generative AI applications. This launch enables you to use Step Functions to orchestrate interactions with foundation models and add in steps for human intervention, without needing to write or maintain integration code. You can leverage Step Functions Workflow Studio to visually develop, inspect, and audit user defined generative AI workflows with ease.
Step Functions provides integrations with two Bedrock APIs. The first is the InvokeModel API integration, which enables you to orchestrate interactions with foundation models, including the ability to read input from and write output to Amazon S3. The second is the CreateModelCustomizationJob API integration which supports the Run a Job (.sync) integration pattern, meaning it waits for the asynchronous job to complete before progressing to the next step in your workflow, making it simpler to integrate model fine-tuning into your data processing pipelines.
To get started, you can use the new “Perform prompt-chaining with Bedrock” Sample Project for Step Functions in the AWS Console or read the Launch Blog. To learn more, please see the AWS Step Functions Developer Guide.