SkillMatched

SkillMatched is a backend-focused job recommendation platform that matches users to relevant opportunities using cosine similarity scoring across 10,000+ simulated job vectors, with Redis caching delivering sub-3ms p99 response times.

---

Features

• Redis-cached job matching: 99.5% latency reduction (600ms → <3ms) on match queries
• Cosine similarity scoring engine across 10,000+ simulated job vectors
• Weekly automated batch dispatch via Node-Cron processing 1,000+ user profiles
• Stateless JWT authentication with role-protected REST API routes
• Containerized deployment via Docker with CI pipeline via GitHub Actions
• Structured error handling across all API endpoints

---

Tech Stack

Frontend

• Next.js / React
• TypeScript, Tailwind CSS

Backend

• Node.js
• Express.js

Database & Storage

• PostgreSQL via Neon (primary database)
• Supabase Storage (resume file storage)
• Redis (caching layer)

Authentication & Security

• JSON Web Tokens (JWT)
• Google OAuth
• bcrypt for password hashing

---

Project Structure

SkillsMatched/
│
├── frontend/
│   ├── components/
│   ├── app/
│   │   └── page.tsx
│
├── backend/
│   ├── routes/
│   ├── middleware/
│   └── server.js
│
└── README.md

Installation & Setup

1. Clone the Repository

git clone https://github.com/Crystlfly/SkillsMatched.git
cd SkillsMatched

2. Backend Setup

cd backend
npm install
npm run dev

Create a .env file in the backend directory:

DATABASE_URL=your_postgres_password
JWT_SECRET=your_jwt_secret

SUPABASE_URL=your_supabase_connection_url
SUPABASE_SERVICE_ROLE=your_service_role

GOOGLE_CLIENT_ID=your_client_id
GOOGLE_CLIENT_SECRET=your_client_secret

EMAIL_USER=your_email_id
EMAIL_PASS=your_email_pass

FRONTEND_URL=http://localhost:3000
NEXT_PUBLIC_BACKEND_URL=http://localhost:5000

REDIS_HOST=your_redis_host
REDIS_PORT=your_redis_port
REDIS_USERNAME=your_redis_username
REDIS_PASSWORD=your_redis_password

3. Frontend Setup

cd ../frontend
npm install
npm run dev

How It Works

Onboarding: User signs up via email or Google OAuth and sets job preferences. Matching: The Node.js backend computes cosine similarity scores across 10,000+ simulated job vectors against the user's preference vector. Caching: Results are cached in Redis — repeated queries are served in under 3ms without recomputation. Dispatch: Node-Cron batches weekly recommendation emails across 1,000+ user profiles.

Use Cases

• Finding technical collaborators for open-source or side projects.
• Skill-based networking within academic or professional circles.
• Pairing mentors with mentees based on specific technology stacks.

Future Enhancements

• Worker thread pool for offloading cosine similarity computation off the main thread
• Real-time job alert notifications

---

If you like this project, please consider giving it a star!