A company builds an internal cost-estimation tool using a SageMaker model. Users upload high-resolution images; the model must process each image, predict the object's cost, and notify the user when processing completes. Which design meets these requirements?

The correct design uses Amazon S3 for image storage, SageMaker asynchronous inference for processing, and Amazon SNS for notifications. S3 is the standard, scalable, and cost-effective storage for large objects like high-resolution images. SageMaker asynchronous inference is ideal for long-running inferences (like processing high-resolution images) where immediate responses aren't needed, and it can handle large payloads. Amazon SNS is a publish/subscribe messaging service well-suited for notifying multiple subscribers (users or other systems) about job completion. Batch transform is less suitable because it's designed for processing entire datasets offline, not for individual, on-demand image processing with user-specific notifications. Amazon EFS is a file system primarily for EC2 instances and is generally more expensive and less scalable for object storage compared to S3, especially for a web application.