📄 Applicant Tracking System (ATS) Project

Applicant Tracking System (ATS) application for automating the application process, CV/Resume text extraction, and matching candidate compatibility with job posting qualifications using NLP (Natural Language Processing) models.

---

🚀 Overview

This ATS system is a complete monorepo platform that integrates an interactive frontend, a secure backend API, a PostgreSQL relational database, and a Python-based AI processing microservice. This system simplifies the recruitment process through:

• Job Posting Management (Job Board): HR can publish, update, and delete job postings along with their qualifications.
• Job Application: Candidates can apply to job openings by uploading a resume (PDF). The system automatically extracts its text, allowing candidates to apply for jobs matching their qualifications.
• AI-Powered CV Scoring: Uses Sentence-Transformer & SpaCy models to measure the semantic similarity between the candidate's profile/CV and the job qualifications, generating an instant match score (0-100%).

---

🛠️ Tech Stack & Architecture

The system is designed using a monorepo architecture divided into 3 main services:

1. Client (Frontend)
   • Technology: React (Vite), JavaScript, Tailwind CSS, React Router DOM.
   • Function: Provides an interactive dashboard interface for Jobseekers, HR, and Administrators.
2. Server (Backend API)
   • Technology: Node.js, Express.js, PostgreSQL (pg), node-pg-migrate, JWT Authentication.
   • Function: Provides RESTful APIs, manages role-based access control (RBAC), handles CV file uploads (Multer), and coordinates calls to the AI Microservice.
3. AI Service (AI Microservice)
   • Technology: Python (Flask), pdfplumber (for PDF extraction), sentence-transformers (all-MiniLM-L6-v2 model for similarity), spaCy (for custom NER on IT SKILL entities).
   • Function: Securely extracts text from CV files and evaluates the match between the candidate's CV and the job vacancy using Cosine Similarity.
4. Infrastructure
   • Technology: Docker & Docker Compose.
   • Function: Automatically orchestrates containerization for the database, backend, frontend, and AI service within a secure local network.

---

📁 Monorepo Directory Layout

.
├── client/                 # React Frontend Application (Port 3000)
│   ├── src/
│   │   ├── components/
│   │   ├── constants/
│   │   ├── context/
│   │   ├── pages/
│   │   ├── services/
│   │   ├── shared/
│   │   ├── utils/
│   │   ├── App.jsx
│   │   ├── index.css
│   │   └── main.jsx
│   ├── Dockerfile
│   └── .env.example
├── server/                 # Express Backend API (Port 5000)
│   ├── src/
│   │   ├── config/
│   │   ├── controllers/
│   │   ├── middleware/
│   │   ├── models/
│   │   ├── routes/
│   │   ├── scripts/
│   │   ├── services/
│   │   ├── utils/
│   │   ├── validators/
│   │   └── index.js
│   ├── migrations/
│   ├── Dockerfile
│   └── .env.example
├── ai-service/             # Flask AI Service (Port 5001 - Internal)
│   ├── app/
│   │   ├── routes/
│   │   ├── services/
│   │   ├── utils/
│   │   ├── config.py
│   │   └── logging_config.py
│   ├── main.py
│   ├── Dockerfile
│   └── .env.example
├── erd.mermaid             # Entity-Relationship Diagram (Mermaid Format)
├── docker-compose.yml      # Docker container orchestration configuration
└── .env.example            # Global environment variable template

---

⚙️ Core API Endpoints

1. Backend Service (server) - Port 5000

Method	Endpoint	Access Level (Role)	Description
POST	/api/auth/register	Public	Register a new user account for jobseekers
POST	/api/auth/login	Public	Authenticate user & return JWT Token
POST	/api/auth/change-password	Authenticated (All Roles)	Change password of the active account
POST	/api/admin/users/hr	Admin Only	Create a new account specifically for HR staff
GET	/api/jobs	Public / Authenticated	View list of active job postings
POST	/api/jobs	Admin / HR	Create a new job posting
GET	/api/jobs/:id	Public / Authenticated	View details of a specific job posting
PUT	/api/jobs/:id	Admin / HR	Update information of a job posting
DELETE	/api/jobs/:id	Admin / HR	Delete a job posting
POST	/api/jobs/:jobId/apply	Jobseeker Only	Submit application by attaching CV file
GET	/api/jobs/:jobId/candidates	Admin / HR	View list of candidates for a specific job
GET	/api/jobs/:jobId/candidates/:userId	Admin / HR	View candidate profile details & AI compatibility score
GET	/api/jobs/:jobId/candidates/:userId/cv/download	Admin / HR	Download original candidate CV file
GET	/api/health	Public	Check backend health status
GET	/api/health/ai	Public	Check connection from Backend to AI Service

2. AI Service (ai-service) - Port 5001 (Internal Only)

Method	Endpoint	Description	Input Format / Payload
POST	/extract-text	Extract raw text from PDF file	Multipart Form-Data (file: PDF)
POST	/score-cv-job	Calculate match score between CV vs Job posting	JSON: application_id, extracted_text_cv, job_info
GET	/health	Check AI service operational status	None

---

How to Run the Application

Using Docker Compose

The fastest way to spin up the entire system along with its configuration:

1. Ensure environment variables (.env) are correctly configured.
2. Ensure Docker & Docker Compose are running on your system.
3. Build & run all containers in the background:

   docker compose up --build -d

4. Check container status:

   docker compose ps