// the build story
A production-grade platform that turns scattered campus opportunities – internships, research, hackathons – into one structured, searchable pipeline.
In my third year I noticed a gap everyone lived with but nobody named: faculty and departments held opportunities – internships, research openings, hackathons – that rarely reached the students they were meant for. What did circulate moved through WhatsApp forwards, expired Google Forms, and word of mouth, and the good ones were usually found too late.
Nobody asked me to fix this. The observation was mine, and so was the itch: this wasn't a content problem, it was an infrastructure problem. The opportunities existed; there was just no reliable, structured pipe between the people who had them and the people who needed them.
I designed Opportunity Quest as a single source of truth: a platform where opportunities carry structured metadata – type, deadline, eligibility – and students query with real filters instead of scroll-and-pray. The system was designed before the code: entity relationships, the route map, and the auth flows all existed on paper first.
From day one I treated it as a production system, not a college project. The data model was locked before the first route was written, role-based access (students vs. posters vs. admins) was planned into the schema up front, and security was part of the architecture – threat-modeled around the auth surface – rather than a final-week patch.
The result is a MERN application with a clean RESTful API layer. Authentication uses JWTs delivered via httpOnly cookies, keeping tokens out of localStorage and out of reach of XSS. CORS is locked to known origins, secrets live in environment variables, and rate limiting protects the auth endpoints from brute-force attempts.
On the frontend, React handles the discovery flow: browsing, filtering, and viewing opportunity detail pages, with auth state managed cleanly across protected routes. The API is organized by resource with consistent error handling, making new opportunity types cheap to add.
A deployable, security-hardened platform that covers the full engineering lifecycle solo: schema design, API architecture, auth, and frontend state, plus the production concerns (rate limiting, cookie security, CORS policy) that separate a demo from a system. It's the kind of infrastructure campuses usually buy, built from a gap I noticed and refused to ignore.
The codebase is structured so the next features – notifications, saved searches, posting workflows – slot into existing patterns instead of requiring rewrites. That's deliberate: the platform was engineered for the roadmap, not just the demo.
Storing tokens in localStorage is the path of least resistance, but it leaves them readable by any injected script. httpOnly cookies cost more setup (CSRF consideration, CORS credentials) but remove an entire class of XSS token theft.
Roles were modeled before the first route was written, so authorization checks are a middleware concern, not conditionals scattered through controllers.
Login and signup are the highest-value attack surface. Express rate limiting keeps brute-force attempts cheap to absorb without touching legitimate traffic.
For a resource-oriented domain (opportunities, users, applications), REST keeps the mental model simple and caching trivial. GraphQL's flexibility wasn't worth its complexity at this scale.