Back to problems

Design and explain robust web APIs for ML inference

System Design · NVIDIA · Medium

Design a RESTful HTTP interface for serving predictions from image models. The interface should let callers submit image payloads, track asynchronous work, and fetch inference results. Your specification must cover the following: Route design: define the HTTP endpoints and verbs for creating an inference job, checking status, and retrieving results; describe how API versions are embedded in paths or headers and how idempotent retries are supported. Message formats: specify…

Checking your access…