API Design in System Design Interviews w/ Meta Staff Engineer

In summary, the most interesting aspect of this transcript is the strategic framework for selecting API protocols based on context: defaulting to REST for external client-facing systems, leveraging RPC/gRPC for high-performance internal microservice communication, and utilizing GraphQL for flexible, client-driven data fetching, all while emphasizing the critical interview skill of maintaining brevity and avoiding unnecessary technical deep dives.

• Prioritize brevity in interviews, spending no more than 5 minutes on API design to avoid derailing the broader system architecture discussion.
• REST is the default choice for external APIs, utilizing standard HTTP methods (GET, POST, PUT, PATCH, DELETE) and plural nouns for resource paths.
• GraphQL solves data overfetching and underfetching by allowing clients to request exact fields in a single query, though it requires batching strategies to avoid the N+1 database problem.
• RPC and gRPC are superior for internal microservice-to-microservice communication due to binary protocols, reduced HTTP overhead, and significantly faster data transfer.
• Implement pagination for large datasets, using offset-based methods for simplicity and cursor-based methods to maintain consistency under heavy write loads.
• Enforce security via authentication tokens (JWTs or session IDs) in HTTP headers rather than trusting client-provided user IDs in request bodies.