🤖 Smart Job Hunter

An AI-powered, full-stack job application automation platform that uses LLMs and browser automation to search, evaluate, and apply to jobs on your behalf.

📋 Table of Contents

🎯 Overview

Smart Job Hunter is a full-stack platform that automates the job application process on Naukri.com. Upload your resume, set your preferences, and let the AI agent handle the rest — from searching for relevant roles, to scoring job matches against your skills, to clicking "Apply" on your behalf.

The system combines:

LLM-powered resume parsing (Groq API + Llama) to extract skills, experience, and role from your PDF resume.
Hybrid browser automation (Playwright) using both deterministic scripts and LLM-driven decision-making for navigating complex, dynamic web pages.
Intelligent job matching with skill synonym expansion, role allowlists, and configurable match thresholds.

🏗 Architecture

┌──────────────────────────────────────────────────────────────┐
│                     React Frontend (Vite)                    │
│  ┌─────────────┬──────────────┬─────────────┬──────────────┐ │
│  │ResumeUpload │JobPreferences│ AppLogs     │ Settings     │ │
│  └──────┬──────┴──────┬───────┴──────┬──────┴──────┬───────┘ │
└─────────┼─────────────┼──────────────┼─────────────┼─────────┘
          │  REST API   │              │             │
┌─────────▼─────────────▼──────────────▼─────────────▼─────────┐
│                 Spring Boot Backend (Java 21)                 │
│  ┌───────────────┬─────────────────┬───────────────────────┐  │
│  │ResumeController│AgentController │SessionController      │  │
│  │  (PDF → LLM)   │ (Start Agent) │ (Naukri Login/Session)│  │
│  └───────┬────────┴───────┬────────┴───────────┬──────────┘  │
│          │                │                    │              │
│  ┌───────▼────────────────▼────────────────────▼──────────┐  │
│  │              PlaywrightService                          │  │
│  │   Spawns Node.js scripts as child processes             │  │
│  └────────────────────────┬───────────────────────────────┘  │
└───────────────────────────┼──────────────────────────────────┘
                            │
┌───────────────────────────▼──────────────────────────────────┐
│              Playwright Automation Scripts (Node.js)          │
│  ┌──────────┬──────────┬──────────────┬────────────────────┐ │
│  │ login.js │ apply.js │  agent.js    │   llm-brain.js     │ │
│  │ (Manual  │(Legacy   │(Hybrid v5.0)│  (Groq LLM Client) │ │
│  │  Login)  │ Apply)   │ Deterministic│  decideAction()   │ │
│  │          │          │ + LLM Agent  │  parseAction()    │ │
│  └──────────┴──────────┴──────────────┴────────────────────┘ │
└──────────────────────────────────────────────────────────────┘
                            │
                    ┌───────▼───────┐
                    │   MongoDB     │
                    │ (Sessions,    │
                    │  App Logs)    │
                    └───────────────┘

✨ Key Features

Feature	Description
📄 AI Resume Parsing	Upload a PDF resume → PDFBox extracts text → Groq LLM extracts name, skills, experience, and detected role
🔐 Session Management	Playwright opens a real Chromium browser for manual Naukri.com login; session is persisted and reused across runs
🤖 Hybrid Automation Agent	Deterministic logic for search/navigation + LLM-driven decisions for complex UI interactions (modals, questionnaires)
🎯 Smart Job Matching	Skill synonym expansion (e.g., "JS" → "JavaScript"), role allowlists, and configurable match threshold (20-100%)
🛡 Anti-Detection	Stealth scripts, randomized viewports, humanized mouse movements (Bézier curves), and natural typing delays
📊 Application Logs	Real-time dashboard showing applied/skipped/failed jobs with match scores and status details
⚙️ Configurable Settings	Daily application limit, match threshold, and skill overrides — all adjustable from the UI
🔄 Session Expiry Handling	Auto-detects expired sessions and prompts reconnection

🛠 Tech Stack

Backend

Technology	Purpose
Java 21	Primary backend language
Spring Boot 3.5	REST API framework
Spring WebFlux	Non-blocking HTTP client for Groq API calls
MongoDB	Stores user sessions and application logs
Apache PDFBox 3.x	PDF text extraction from resumes
Lombok	Reduces boilerplate (DTOs, models)

Frontend

Technology	Purpose
React 19	UI library
Vite 8	Build tool & dev server
Lucide React	Icon library
Vanilla CSS	Custom glassmorphism dark theme

Automation

Technology	Purpose
Playwright	Browser automation (Chromium)
Groq API	LLM inference (Llama 4 Scout)
Node.js	Script runtime for automation agents

📁 Project Structure

smart-job-hunter/
├── Frontend/
│   └── vite-project/
│       ├── src/
│       │   ├── App.jsx                # Main app with tab navigation
│       │   ├── App.css                # Glassmorphism dark theme
│       │   ├── index.css              # Global styles
│       │   ├── components/
│       │   │   ├── ResumeUpload.jsx   # PDF upload + AI parsing
│       │   │   ├── JobPreferences.jsx # Role, location, skills + Naukri connect
│       │   │   ├── ApplicationLogs.jsx# Real-time job application logs
│       │   │   ├── SettingsPanel.jsx  # Daily limit & threshold config
│       │   │   └── OtpModal.jsx       # OTP handling modal
│       │   └── assets/
│       └── package.json
│
├── job-hunter/                         # Spring Boot Backend
│   ├── src/main/java/com/omkar/jobhunter/
│   │   ├── Application.java           # Spring Boot entry point
│   │   ├── config/
│   │   │   └── WebConfig.java         # CORS configuration
│   │   ├── controller/
│   │   │   ├── ResumeController.java  # POST /api/resume/parse
│   │   │   ├── AgentController.java   # POST /api/agent/start
│   │   │   └── SessionController.java # POST /api/session/connect/naukri
│   │   ├── dto/
│   │   │   ├── AgentRequest.java      # Agent start request DTO
│   │   │   ├── AgentResponse.java     # Agent response DTO
│   │   │   └── OtpSubmitRequest.java  # OTP submission DTO
│   │   ├── model/
│   │   │   ├── ApplicationLog.java    # MongoDB document for job logs
│   │   │   └── UserSession.java       # MongoDB document for sessions
│   │   ├── repository/
│   │   │   ├── ApplicationLogRepository.java
│   │   │   └── UserSessionRepository.java
│   │   └── service/
│   │       └── PlaywrightService.java # Spawns Node.js automation scripts
│   ├── scripts/                       # Playwright Automation
│   │   ├── agent.js                   # Hybrid v5.0 — main automation agent
│   │   ├── apply.js                   # Legacy apply script
│   │   ├── login.js                   # Manual login via Playwright
│   │   ├── llm-brain.js              # Groq LLM client (decideAction)
│   │   └── package.json
│   └── pom.xml
│
├── .gitignore
└── README.md

📋 Prerequisites

Java 21 (JDK)
Node.js 18+ and npm
MongoDB (running locally on port 27017)
Groq API Key — Get one free at groq.com
Maven (or use the included mvnw wrapper)

🚀 Getting Started

1. Clone the repository

git clone https://github.com/ItsOmkara/smart-job-hunter.git
cd smart-job-hunter

2. Configure the backend

Create job-hunter/src/main/resources/application.properties:

spring.data.mongodb.uri=mongodb://localhost:27017/jobhunter
spring.data.mongodb.database=jobhunter

# Playwright config
playwright.scripts.dir=scripts
playwright.browser.datadir=browser-data

# Groq API (for resume parsing + LLM agent)
groq.api.key=YOUR_GROQ_API_KEY_HERE
groq.model=meta-llama/llama-4-scout-17b-16e-instruct

3. Install & run the backend

cd job-hunter
./mvnw spring-boot:run

The API server starts at http://localhost:8080.

4. Install Playwright browsers

cd job-hunter/scripts
npm install
npx playwright install chromium

5. Install & run the frontend

cd Frontend/vite-project
npm install
npm run dev

The frontend starts at http://localhost:5173.

6. Use the app

Upload Resume — Drop your PDF on the dashboard. The AI extracts your skills and role automatically.
Connect Naukri — Click "Connect Naukri" to open a Chromium window. Log in manually; the session is captured automatically.
Set Preferences — Choose target role, location(s), experience level, and adjust the match threshold.
Start Applying — Click "Start Autonomous Applying" and watch the logs roll in!

📡 API Reference

Resume

Method	Endpoint	Description
`POST`	`/api/resume/parse`	Upload PDF resume → returns extracted JSON (name, skills, role, experience)

Session

Method	Endpoint	Description
`POST`	`/api/session/connect/naukri`	Opens Chromium for manual login; persists session
`GET`	`/api/session/status/naukri`	Check if an active Naukri session exists

Agent

Method	Endpoint	Description
`POST`	`/api/agent/start`	Start the automation agent with role, location, skills, limits
`DELETE`	`/api/agent/logs`	Clear all application logs

`POST /api/agent/start` — Request Body

{
  "role": "Full Stack Developer",
  "location": "Pune, Bangalore",
  "experience": "entry",
  "skills": "React, Java, Spring Boot, MongoDB",
  "dailyLimit": 5,
  "matchThreshold": 25
}

⚙️ How It Works

1. Resume Parsing Pipeline

PDF Upload → PDFBox text extraction → Groq LLM (Llama 3.3 70B) → JSON { name, skills, role, experience }

2. Session Management

Playwright launches a persistent Chromium browser (with userDataDir)
User logs in manually → cookies/session are saved to disk
Subsequent agent runs reuse the same browser profile — no re-login needed

3. Hybrid Agent (v5.0)

The agent combines two strategies:

Phase	Strategy	What It Does
Search & Navigate	Deterministic	Constructs Naukri.com search URLs, handles pagination
Job Extraction	Deterministic	Parses job cards, extracts title/company/skills
Role Filtering	Deterministic	Uses role allowlist + exclusion list to filter irrelevant jobs
Skill Matching	Deterministic	Expands skill synonyms, calculates match score
Apply Flow	LLM-Driven	Handles dynamic modals, questionnaires, and complex UI states via Groq API

4. Anti-Detection

Stealth scripts (navigator overrides, plugin simulation)
Randomized viewport sizes
Humanized mouse movements using Bézier curves
Natural typing delays (jitter per keystroke)
Watchdog timer for stuck-state recovery

🔧 Configuration

Application Settings (via UI → Settings Panel)

Setting	Default	Description
Daily Limit	`5`	Max applications per agent run
Match Threshold	`25%`	Minimum skill match score to apply

Agent Configuration (via `agent.js`)

Config	Description
`ROLE_ALLOWLIST`	Map of target roles → acceptable title variants
`EXCLUSIONS`	Job titles to always skip (sales, HR, etc.)
`SKILL_SYNONYMS`	Maps skill names to equivalents (e.g., `react` → `reactjs`)
`WATCHDOG_TIMEOUT`	Auto-exit if no progress for 3 minutes

🤝 Contributing

Fork the repository
Create a feature branch: git checkout -b feature/amazing-feature
Commit your changes: git commit -m 'Add amazing feature'
Push to the branch: git push origin feature/amazing-feature
Open a Pull Request

📄 License

This project is open source and available under the MIT License.

Built with ❤️ by Omkar

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
Frontend/vite-project		Frontend/vite-project
job-hunter		job-hunter
.gitignore		.gitignore
README.md		README.md

Folders and files

Latest commit

History

Repository files navigation

🤖 Smart Job Hunter

📋 Table of Contents

🎯 Overview

🏗 Architecture

✨ Key Features

🛠 Tech Stack

Backend

Frontend

Automation

📁 Project Structure

📋 Prerequisites

🚀 Getting Started

1. Clone the repository

2. Configure the backend

3. Install & run the backend

4. Install Playwright browsers

5. Install & run the frontend

6. Use the app

📡 API Reference

Resume

Session

Agent

POST /api/agent/start — Request Body

⚙️ How It Works

1. Resume Parsing Pipeline

2. Session Management

3. Hybrid Agent (v5.0)

4. Anti-Detection

🔧 Configuration

Application Settings (via UI → Settings Panel)

Agent Configuration (via agent.js)

🤝 Contributing

📄 License

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

`POST /api/agent/start` — Request Body

Agent Configuration (via `agent.js`)

Packages