Hire elite TypeScript developers.
Type-safe full-stack, AI-enhanced.

Expected answer

Both define object shapes, but interfaces support declaration merging and extending with `extends`, while types support unions, intersections, mapped types, and conditional types. Use interfaces for public API contracts (they merge across declarations) and types for complex type computations (unions, mapped types, utility types). In practice, for most object shapes they're interchangeable.

Expected answer

Use tRPC for end-to-end type safety without code generation. Define procedures on the server with input validation (Zod), and the client automatically infers types. Alternative: GraphQL with codegen (graphql-codegen). Both eliminate the contract mismatch between frontend and backend. With tRPC, changing a server response type immediately shows errors in the client — zero runtime surprises.

Expected answer

Discriminated unions use a common literal property (discriminant) to narrow types. Example: type State = { status: "loading" } | { status: "success"; data: User } | { status: "error"; error: Error }. When you check state.status, TypeScript narrows the type so you can safely access state.data only when status is "success". This eliminates impossible states and makes state machines type-safe.

Expected answer

Conditional types follow the pattern: T extends U ? X : Y. Practical example: type ApiResponse<T> = T extends Array<infer U> ? { items: U[]; total: number } : { item: T }. This creates different response shapes for arrays vs. single items. They're powerful for library types — e.g., Prisma uses conditional types extensively to infer query result types based on the `select` and `include` options you pass.

Expected answer

TypeScript types are erased at runtime, so you need runtime validation at system boundaries (API inputs, form data, env vars). Use Zod: define a schema, parse input, and infer the TypeScript type from the schema. const UserSchema = z.object({ name: z.string(), email: z.string().email() }); type User = z.infer<typeof UserSchema>. This gives you a single source of truth for both runtime validation and compile-time types.