Topic overview
Study path and query parameters, request/response bodies via Pydantic models, dependency injection for shared logic like auth and DB sessions, async endpoints, background tasks, and the automatic OpenAPI docs generated from your type hints.
Core concepts
Path operations and routing, Pydantic-based request validation and serialization, Depends() for reusable dependencies, async def vs plain def endpoints, background tasks for fire-and-forget work, and middleware for cross-cutting concerns.
Why it matters
FastAPI is the default choice for serving ML/AI models and building typed backend APIs quickly — validation and docs come for free from the same type hints you'd write anyway, which is why it's replaced Flask in most new Python services.
Interview relevance
Expect questions on when async def actually helps (I/O-bound work) versus when a blocking call inside it stalls the whole event loop, how dependency injection differs from just calling a function directly, and how request validation errors are surfaced to clients.