Skip to content

Interview Cheat Sheet: Backend/Fullstack Engineer

Night-before review. Key patterns, talking points, and terms for the NestJS/GraphQL/PostgreSQL stack. Bridged from your JVM/Scala/Rust background.

Your Unique Angles

Their need Your strength Talking point
NestJS architecture Spring/Scala DI expertise "Same DI patterns, different syntax -- I'll be productive fast"
GraphQL Type system expertise (Scala 3, Rust) "I think in types; GraphQL schema design is natural"
PostgreSQL Database experience from JVM projects "I know EXPLAIN ANALYZE and index tuning from production"
Scaling Akka actors, distributed systems MSc "BullMQ jobs = Akka actors with persistence"
Auth/security Unix security model, SSH expertise "Security is in my DNA from Linux systems work"
Mentoring 5 years daily coding, architecture mindset "I lead by writing clear, reviewable code"

JVM → NestJS Translation Table

JVM/Scala NestJS Notes
@Autowired / ZIO Layer @Injectable() + constructor DI Same concept, decorators instead of annotations
@Configuration @Module() Module = bounded context
@PreAuthorize @UseGuards(RolesGuard) Guards = authorization layer
AOP @Around @UseInterceptors() Interceptors wrap handler execution
Bean Validation class-validator + ValidationPipe Decorators on DTO classes
Akka Actor BullMQ Job + Processor Queue = mailbox, processor = actor
Sangria ObjectType @ObjectType() Code-first GraphQL
Slick / JPA Entity TypeORM @Entity() Decorator-based ORM

Key Patterns to Know

NestJS Execution Pipeline

Request → Middleware → Guard → Interceptor(pre) → Pipe → Handler → Interceptor(post) → Response
                                                                          ↓ on error
                                                                   Exception Filter

GraphQL Subscriptions (real-time VNC events)

Client ←WebSocket→ NestJS Gateway ←Redis PubSub→ Other NestJS instances

Horizontal Scaling

Load Balancer → NestJS Node 1 ──┐
             → NestJS Node 2 ──┤── Redis (sessions, PubSub, rate limits, BullMQ)
             → NestJS Node N ──┘── PostgreSQL + PgBouncer (connection pooling)

Questions They Might Ask

  1. "Design the backend for our AI agent session management." → NestJS service creates Kubernetes pod via BullMQ job. WebSocket gateway streams VNC frames + AI actions. GraphQL subscriptions push status. PostgreSQL stores session log (JSONB). RLS isolates tenants.

  2. "How would you handle 10,000 concurrent WebSocket connections?" → Socket.io with Redis adapter (broadcasts across nodes). Stateless NestJS behind round-robin LB. Redis for session state. Health checks via terminus.

  3. "Your SQL queries are slow. How do you diagnose?" → EXPLAIN (ANALYZE, BUFFERS). Look for Seq Scans on large tables (missing index), Nested Loops with high actual rows (N+1), Sort with external merge (insufficient work_mem). Check pg_stat_user_tables for dead tuples (vacuum needed).

  4. "How do you prevent the N+1 problem in GraphQL?" → DataLoader: batch .load(id) calls within one tick into a single WHERE id IN (...). Per-request scoping via GraphQL context factory. 51 queries → 2 queries.

Numbers to Know

  • NestJS 11: Express v5 + Fastify v5, breaking route syntax changes
  • PostgreSQL 17: incremental backup (78min → 4min recovery), JSON_TABLE
  • PgBouncer: ~2KB per connection vs PostgreSQL's ~5-10MB
  • BullMQ: Redis-backed, supports retries, priorities, rate limiting, progress
  • Argon2id: OWASP recommended, 19 MiB memory, 2 iterations minimum
  • TypeORM: maintenance mode for new features; Drizzle gaining ground
  • GraphQL subscriptions: graphql-ws protocol (not subscriptions-transport-ws)