SageMaker Async Inference Now Supports Inline Payloads

Amazon has announced a significant update to its SageMaker AI Async Inference service, introducing support for inline request payloads. Previously, users were required to upload their input data to an Amazon S3 bucket before initiating an asynchronous inference request. With this new feature, data can now be directly included in the request body of the InvokeEndpointAsync API call. The update aims to reduce the complexity and overhead associated with managing temporary S3 storage for inference inputs. By allowing inline payloads, the process becomes more efficient, particularly for smaller data inputs or scenarios where direct API interaction is preferred, ultimately accelerating the deployment and testing of AI models on SageMaker.