AWS published a post on end-to-end encrypted ML inference with Amazon SageMaker AI and fully homomorphic encryption. The guide builds on earlier work around FHE-based inference and explains how teams can protect sensitive inputs during real-time model serving workflows.