GraphQL API design

GraphQL API interviews focus on schema design, resolver performance, batching, federation, and protecting the server from expensive queries—not replacing REST by default.

Subtopics: schema and operations, resolvers with N+1 and DataLoader, and federation with query complexity limits.

On interviews: sketch a schema for a nested read, explain how you prevent N+1, and describe depth/cost limits at the gateway.

Common pitfalls: resolvers that hit the database per field, missing authorization on nested fields, and public GraphQL without persisted operations or complexity rules.

Checklist:

• Model types and nullability deliberately.
• Batch per-request data access with DataLoader.
• Enforce auth at field or object boundaries.
• Limit depth, cost, and introspection in production.