Crackd
AI placement & roadmap platform Live
01 — AI placement & roadmap platform
Crackd
A MERN platform for placement preparation. Students explore companies, share interview experiences, generate and track weekly study roadmaps, and ask an AI assistant questions grounded in real interview data from their seniors.
18REST API endpoints
10 minRedis cache TTL
CILint, build & tests per PR
Key architectural highlights
- RAG preparation assistant: Interview experiences are embedded with Gemini
(
gemini-embedding-001) and retrieved by cosine similarity, so answers cite real experiences — and the assistant says so when it has no matching data. - AI roadmaps with a safe fallback: Gemini generates a week-by-week plan from the target role, company, timeline, hours per day and strong/weak topics; if generation fails, the API returns a template roadmap instead.
- Progress tracking: One active roadmap per user with checklist updates, a readiness score and study streaks.
- Auth & caching: Registration limited to
@nie.ac.inaddresses, bcrypt password hashing and JWT-protected routes; the company list is cached in Redis for 10 minutes, and the app keeps running if Redis is unavailable. - Tested & automated:
node:testunit tests for the auth and college-email middleware plus Jest + Supertest API tests, run by GitHub Actions on every push and pull request.
