A company needs to host a custom ML model to perform forecast analysis.

SageMaker Serverless Inference is ideal for workloads with predictable, intermittent demand. By enabling provisioned concurrency, the model can handle multiple invocations quickly during the high-demand 2-hour period. AWS manages the underlying infrastructure and scaling, ensuring the solution meets performance requirements with minimal operational overhead. This approach is cost-effective since it scales down when not in use.