Career roadmap
Full-Stack Engineer
Own a feature from database to pixel. The default hire at startups and the most common job title in the market.
Before you start Full-Stack
- JavaScript and one backend language
- HTML, CSS and basic SQL
- Git, GitHub, and a deployed project
One vertical slice
Ship one feature through every layer before widening. Depth first, then breadth.
The shared language across both halves of the stack, and the one interviewers probe hardest.
Ch 1 — Closures- Closures, promises, async/await
- TypeScript generics and narrowing
- Module systems and bundling
- Error handling across async boundaries
Component model, state and effects — enough to build product UI without fighting the framework.
- Components, props and composition
- State, effects and the dependency array
- Lists, keys and conditional rendering
- Routing and layouts
REST, validation and errors. The contract between your two halves.
- REST design and status codes
- Request validation and error shapes
- Middleware, logging and request context
- OpenAPI documentation
Schema design and migrations, plus enough SQL to escape the ORM when it fails you.
- Schema design and relationships
- Migrations and rollback
- ORM query patterns and their SQL output
- Indexes and the N+1 problem
Sessions or tokens, on both sides of the wire. The most common full-stack take-home.
- Session cookies vs JWT
- Password hashing and reset flows
- Protected routes on client and server
- OAuth social login
BuildAn authenticated CRUD app: database, API, UI, deployed with a real domain and HTTPS.
Production quality
The difference between a demo and something a team would let you deploy on a Friday.
Full-stack testing means knowing what to test at which layer, and not testing everything twice.
- Component tests with Testing Library
- API integration tests against a real database
- One end-to-end happy path with Playwright
- Test data and fixtures
Being the person who set up the pipeline is a strong signal at startup interviews.
- GitHub Actions: lint, test, build, deploy
- Staging vs production configuration
- Database migrations in the pipeline
- Rollback plan
Knowing something broke before your users tell you.
- Structured logs with request ids
- Error tracking on client and server
- Uptime checks and alerting
- Basic dashboards
XSS, CSRF, injection and secrets. Enough to not be the reason for the breach.
- OWASP Top 10 in practice
- CSRF protection and SameSite cookies
- Input sanitisation and output encoding
- Secrets and environment configuration
The full-stack advantage: you can tell whether the problem is the query or the bundle.
- Core Web Vitals on the frontend
- Query profiling on the backend
- Caching at both layers
- Image and asset optimisation
BuildAdd tests, CI, error tracking and a staging environment to the app from stage 1.
Framework depth
Pick one modern full-stack framework and know it properly, including where it runs your code.
The server/client boundary is the defining full-stack question of the current stack.
- Server components vs client components
- Data loading and caching on the server
- Server actions and mutations
- Streaming and Suspense boundaries
NestJS, Django or Rails — one opinionated framework you can defend in interview.
Ch — Express vs NestJS- Modules, providers and dependency injection
- Guards, interceptors and pipes
- Background jobs and scheduling
- Framework testing utilities
Chat, notifications and live updates come up in almost every product take-home now.
- WebSockets vs SSE vs polling
- Presence and reconnection handling
- Scaling sockets across instances
- Optimistic UI with rollback
Every product needs it, and doing it correctly with presigned URLs is a good signal.
- Presigned uploads to object storage
- Image resizing and CDN delivery
- Virus scanning and content type checks
- Large file and resumable uploads
Webhook handling and idempotency are the parts candidates get wrong.
- Stripe checkout and webhooks
- Idempotency keys and replay safety
- Reconciliation and failed payments
- PCI scope awareness
BuildRebuild your stage 1 app in a server-rendered framework with streaming and server-side data loading.
Architecture and scale
Answer 'what happens when this has a hundred times the traffic' without hand-waving.
Full-stack loops ask product-shaped design questions: design Instagram stories, not design Kafka.
Ch — Scalable APIs- Requirements and scale estimation
- Data model and access patterns
- Caching and CDN strategy
- Identifying the bottleneck
Browser, CDN, application and database. Knowing which layer to fix is the skill.
- HTTP caching and CDN rules
- Redis application cache
- Query result caching and invalidation
- Stale-while-revalidate patterns
Anything slow belongs off the request path. Interviewers check you know which things.
- Job queues and workers
- Scheduled and recurring jobs
- Retries and dead letter handling
- Email, export and report generation
Nearly every B2B product needs it, and getting isolation wrong is a security incident.
- Row-level vs schema-level isolation
- Role and permission modelling
- Tenant-scoped queries by default
- Admin impersonation done safely
In 2026 most product roles expect you to wire an LLM into a feature competently.
Ch — RAG- Calling an LLM API with streaming
- Retrieval over your own data
- Cost, latency and caching
- Evaluating output quality
BuildTake one endpoint to 1,000 requests per second on a load test, and write up what you changed.
Interview and evidence
Full-stack loops mix a take-home, a system design round, and DSA. Prepare all three.
The most common full-stack filter. Scope control and a good README beat extra features.
- Reading the brief for the real requirements
- Scoping to the time limit deliberately
- README: decisions, trade-offs, what you skipped
- Tests that demonstrate judgement
Still a gate at most product companies, even for product-focused roles.
DSA Coding Questions- Arrays, strings and hash maps
- Two pointers and sliding window
- Trees, recursion and BFS/DFS
- Complexity analysis
Increasingly common: review a pull request out loud and say what you would block on.
- Spotting security and correctness issues
- Distinguishing blocking from nit
- Giving feedback that lands
- Reading unfamiliar code quickly
Reviewers spend three minutes. Make the first screen of the README do the work.
- Problem, architecture, trade-offs, screenshots
- Live demo with seeded data
- Setup that works from a clean clone
- Honest limitations section
Startups probe autonomy and judgement more than process. Have stories with numbers.
- Shipping something end to end alone
- A trade-off you made under time pressure
- Working directly with users or stakeholders
- Learning an unfamiliar part of the stack fast
BuildOne flagship product, deployed, with real users if you can get them, and a written architecture document.
Full-Stack tools on your CV
- TypeScript
- React
- Next.js
- Node.js / NestJS
- PostgreSQL
- Prisma
- Redis
- Docker
- Playwright
What Full-Stack employers ask to see
- One deployed product with authentication, payments and real data
- A take-home repo with a README a reviewer can follow in three minutes
- A load test result showing a bottleneck found and fixed
- An architecture document for your flagship project
The broadest job title in software. Startups hire almost exclusively for it, and agencies and product teams treat it as the baseline. Breadth gets you hired; one deep spike gets you levelled.
Content last reviewed 2026-08-31. Guidance only — no institute or paid placement is endorsed anywhere in this book.