Job Queue

Agentic Job Queue is an AI-assisted job-search platform that helps users upload a resume, save job preferences, discover matching jobs, queue suitable opportunities, mark application progress, and save favorite jobs.

The project is built as a full-stack JavaScript application with a React/Vite frontend, an Express backend, MongoDB persistence, JWT authentication, PDF resume parsing, job discovery through JSearch with Remotive fallback, and AI-assisted scoring/generation utilities.

Why this project exists

Most job seekers waste time opening multiple portals, manually filtering jobs, and tracking applications in spreadsheets. Job Queue solves this by creating a focused queue of jobs that match the user's resume and preferences, so the user can review and apply with less noise.

Core features

• User registration and login with JWT authentication
• PDF resume upload and parsing
• Resume-based extraction of skills, projects, experience, and education
• Job preference form for role, location, job type, and minimum match score
• Job discovery using JSearch API, with Remotive fallback when needed
• AI-style match scoring based on skills, projects, and experience
• Queue of matched jobs only
• Favorites section for saving important jobs
• Application status tracking with statuses like queued, applied, interview, rejected, and offer
• Dark mode toggle

Tech stack

Layer	Technology
Frontend	React, Vite, CSS, Framer Motion, Lucide React
Backend	Node.js, Express.js
Database	MongoDB, Mongoose
Authentication	JWT, bcryptjs
Resume parsing	multer, pdf-parse
Job search	JSearch API, Remotive API fallback
AI integration	Google Gemini API

Project structure

Job_Queue/
|-- backend/
|   |-- middleware/
|   |   `-- authMiddleware.js
|   |-- authRoutes.js
|   |-- calculateMatchScore.js
|   |-- generateCoverLetter.js
|   |-- jobApplications.js
|   |-- parseResume.js
|   |-- processJob.js
|   |-- searchJobs.js
|   |-- server.js
|   |-- uploadResumeRoute.js
|   `-- userPreferences.js
|-- frontend/
|   |-- public/
|   |   `-- assets/
|   |-- src/
|   |   |-- App.jsx
|   |   |-- App.css
|   |   |-- auth.js
|   |   `-- components/
|   `-- index.html
|-- package.json
|-- README.md
|-- CONTRIBUTORS.md
`-- LICENSE

Getting started

1. Clone the repository

git clone https://github.com/theashikagupta/Job_Queue.git
cd Job_Queue

2. Install dependencies

npm install

3. Create environment variables

Create a .env file in the project root or inside the backend/ folder:

PORT=9000
MONGODB_URI=your_mongodb_connection_string
MONGODB_DB_NAME=resume_parser
JWT_SECRET=your_strong_jwt_secret

# Optional but recommended for job search
JSEARCH_API_KEY=your_rapidapi_jsearch_key
JSEARCH_API_HOST=jsearch.p.rapidapi.com

# Optional for cover letter generation
GEMINI_API_KEY=your_gemini_api_key
GEMINI_MODEL=gemini-1.5-flash

Important: never commit real API keys, database credentials, or JWT secrets.

4. Run the backend

npm run dev

By default, the server runs on:

http://localhost:9000

5. Run the frontend

In another terminal:

npm run frontend

Open the Vite local URL shown in the terminal.

Main API routes

Method	Route	Purpose
GET	/health	Check server and database status
POST	/api/auth/register	Register a new user
POST	/api/auth/login	Login and receive JWT token
POST	/uploadResume	Upload and parse PDF resume
GET	/api/preferences	Fetch saved job preferences
POST	/api/preferences	Save job preferences
POST	/api/jobs/search	Search jobs, score them, and save results
GET	/api/applications	List user applications
PATCH	/api/applications/:jobId/status	Update application status
PATCH	/api/applications/:jobId/favorite	Toggle favorite job
POST	/api/cover-letter	Generate a cover letter draft

Protected routes require this header:

Authorization: Bearer <token>

How the agentic workflow works

1. The user uploads a resume.
2. The backend parses the PDF and stores structured resume data.
3. The user saves job preferences.
4. The backend searches jobs from external job APIs.
5. Each job is normalized into a common format.
6. The system calculates a match score using resume skills, projects, and experience.
7. Jobs above the selected threshold are queued.
8. The user reviews the queue, favorites jobs, opens apply links, and updates status.

This makes the project agentic because the system does more than display static data. It uses user goals, resume context, external job sources, scoring logic, and saved preferences to actively discover and organize suitable opportunities.

Match score logic

The current scoring system uses weighted matching:

Resume signal	Weight
Skills	50%
Projects	30%
Experience	20%

The score is rule-based and explainable. It can later be upgraded with embeddings or an LLM-based evaluator for deeper semantic matching.

Available scripts

npm run dev       # Start backend server
npm run start     # Start backend server
npm run backend   # Start backend server
npm run frontend  # Start Vite frontend
npm run build     # Build frontend
npm run preview   # Preview production frontend build
npm run check     # Syntax-check backend/helper files

Roadmap

• Smarter resume-job matching with embeddings
• Better dashboard analytics
• Interview reminder workflow
• Resume improvement suggestions
• Safer autofill after explicit user approval
• Deployment setup for frontend and backend

Security notes

• Passwords are hashed using bcryptjs.
• JWT is required for protected user routes.
• Uploaded resumes are temporarily stored and deleted after parsing.
• Secrets must stay in .env and must not be pushed to GitHub.

Author

Built by Ashika Gupta.

License

This project is licensed under the MIT License. See the LICENSE file for details.