What service would you use for scalable machine learning inference?

The most suitable service for scalable machine learning inference is Amazon SageMaker Endpoint. Amazon SageMaker is a fully managed service that provides developers and data scientists with the ability to build, train, and deploy machine learning models quickly.

Specifically, SageMaker Endpoints allow users to deploy their trained models to a scalable and secure environment, making it easy to perform inference on new data. This capability is designed for high availability and can automatically scale up or down based on the incoming request load. This ensures that your machine learning applications can maintain performance and respond quickly to varying traffic levels.

The other options, while important in the AWS ecosystem, do not specifically cater to the unique requirements of machine learning inference. Amazon S3 is primarily a storage solution, Amazon RDS is designed for relational database management, and Amazon CloudFront is a content delivery network designed to deliver web content and media files efficiently. These services do not provide the specialized infrastructure or scalability features necessary for deploying machine learning models for inference at scale.