A company has an image-classification model and wants to deploy it so a web app can get predictions without the company managing any servers. Which solution satisfies this requirement?

Amazon SageMaker Serverless Inference is the correct solution because it provides a fully managed, serverless option for deploying machine learning models, aligning perfectly with the requirement of not managing any servers. It automatically scales compute capacity based on traffic, making it cost-effective for intermittent or unpredictable workloads. Amazon CloudFront is a content delivery network (CDN) and is not designed for deploying or serving ML models. Amazon API Gateway can be used to create an API endpoint for an ML model, but it requires a backend compute service (like AWS Lambda or EC2) to actually host and run the model, which would involve server management. AWS Batch is for running batch computing jobs and is not suitable for real-time, low-latency inference for a web application.