🎯 Automated Resume Relevance Check System

An AI-powered resume evaluation system designed for Innomatics Research Labs to automate resume evaluation against job requirements at scale, providing consistent, fast, and actionable feedback.

📋 Problem Statement

At Innomatics Research Labs, resume evaluation is currently manual, inconsistent, and time-consuming. Every week, the placement team across Hyderabad, Bangalore, Pune, and Delhi NCR receives 18–20 job requirements, with each posting attracting thousands of applications.

Current Challenges:

• ⏰ Delays in shortlisting candidates due to manual review process
• 🔄 Inconsistent judgments as evaluators interpret role requirements differently
• 📈 High workload for placement staff, reducing their ability to focus on interview prep and student guidance
• 🏢 Pressure from hiring companies expecting fast and high-quality shortlists

🎯 Objectives

The Automated Resume Relevance Check System will:

✅ Automate resume evaluation against job requirements at scale
✅ Generate Relevance Score (0–100) for each resume per job role
✅ Highlight gaps such as missing skills, certifications, or projects
✅ Provide fit verdict (High / Medium / Low suitability) to recruiters
✅ Offer personalized improvement feedback to students
✅ Store evaluations in a web-based dashboard accessible to the placement team

🚀 Proposed Solution

We propose building an AI-powered resume evaluation engine that combines rule-based checks with LLM-based semantic understanding:

Core Components:

• Resume Parsing: Extract raw text from PDF/DOCX and standardize formats
• JD Parsing: Extract role title, must-have skills, good-to-have skills, qualifications
• Hybrid Scoring:
  • Hard Match: Keyword & skill check (exact and fuzzy matches)
  • Soft Match: Semantic fit via embeddings + LLM reasoning
• Output Generation: Relevance Score, Missing Elements, and Verdict
• Web Dashboard: Searchable interface for placement team

🏗️ System Architecture

Workflow:

Job Description Upload → Resume Upload → Automated Parsing → Relevance Analysis → Results Storage

Technical Approach:

1. Resume Parsing: PyMuPDF/pdfplumber for PDF, python-docx for DOCX
2. Text Processing: spaCy/NLTK for entity extraction and normalization
3. Hard Matching: TF-IDF, BM25, fuzzy matching for exact keyword matches
4. Semantic Matching: Sentence transformers + Google Gemini for contextual understanding
5. Scoring Engine: Weighted combination of hard and soft matches
6. Web Interface: Streamlit dashboard with role-specific workflows

🛠️ Tech Stack

Backend

• Python 3.11+: Primary programming language
• FastAPI: High-performance backend API framework
• SQLite: Lightweight database for development (PostgreSQL for production)
• LangChain: LLM workflow orchestration
• spaCy: Advanced NLP and entity extraction
• PyMuPDF/pdfplumber: PDF text extraction
• python-docx: DOCX file processing

Frontend

• Streamlit: Interactive web dashboard
• Plotly: Interactive charts and visualizations
• Pandas: Data manipulation and analysis

AI/ML

• Google Gemini: LLM for semantic matching and feedback generation
• Sentence Transformers: Text embeddings for semantic similarity
• TF-IDF/BM25: Keyword matching algorithms
• Fuzzy Matching: Approximate string matching
• ChromaDB: Vector store for embeddings

📦 Installation

Prerequisites

• Python 3.11 or higher
• Google API Key (for Gemini integration)
• Git

Step 1: Clone Repository

git clone <repository-url>
cd code4tech

Step 2: Install Dependencies

# Install Python packages
pip install -r requirements.txt

# For Python 3.12 users (if you encounter compatibility issues)
pip install -r requirements_python312.txt

Step 3: Download spaCy Model

python -m spacy download en_core_web_sm

Step 4: Set Up Environment Variables

# Copy environment template
cp env_example.txt .env

# Edit .env file and add your Google API key
# GOOGLE_API_KEY=your_api_key_here
# DEFAULT_MODEL=gemini-pro
# EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2

Step 5: Get Google API Key

1. Visit Google AI Studio
2. Create a new API key
3. Add it to your .env file

Step 6: Initialize Database

# Database tables will be created automatically on first run
python -c "from src.models.database import create_tables; create_tables()"

🚀 Usage

Quick Start

# Start the application (recommended)
streamlit run streamlit_app.py

This will start:

• Frontend Dashboard: http://localhost:8501
• Backend API: Automatically started in background
• API Documentation: http://localhost:8000/docs

Manual Start (Alternative)

# Terminal 1: Start backend
python main.py

# Terminal 2: Start frontend
streamlit run dashboard.py --server.port 8501

👥 User Workflows

👔 Placement Officer Workflow

1. Access Dashboard: Click "Placement Officer" card
2. Upload Job Description: Upload JD file (PDF/DOCX/TXT)
3. View Parsed Data: System automatically extracts requirements
4. Search & Filter Resumes: Filter by job role, score, verdict
5. Review Results: View evaluation results and rankings
6. Export Data: Download results as CSV

🎓 Student Workflow

1. Access Portal: Click "Student" card
2. Upload Resume: Upload resume file (PDF/DOCX)
3. Browse Jobs: View available job positions
4. Apply for Jobs: Click "Apply for this Job"
5. Get Instant Results: Receive relevance score, verdict, gap analysis
6. Download Report: Get personalized improvement suggestions

📊 Features

🎯 Core Features

• Automated Resume Evaluation: Process PDF/DOCX resumes against job descriptions
• Relevance Scoring: Generate 0-100 relevance scores with High/Medium/Low verdicts
• Gap Analysis: Identify missing skills, certifications, and projects
• Personalized Feedback: Provide improvement suggestions to students
• Real-time Processing: Instant evaluation results (2-5 seconds)
• Scalable Architecture: Handle thousands of resumes weekly

📈 Dashboard Features

• System Overview: Statistics and performance metrics
• Interactive Charts: Verdict distribution and trends
• Search & Filter: Advanced filtering by multiple criteria
• Export Functionality: Download results in CSV format
• Role-based Access: Separate interfaces for officers and students

🔧 Technical Features

• Multi-format Support: PDF, DOCX, DOC, TXT files
• AI-powered Analysis: Google Gemini for semantic understanding
• Robust Parsing: Multiple fallback methods for text extraction
• Error Handling: Comprehensive error management and user feedback
• API Documentation: Auto-generated API docs with Swagger UI

📁 Project Structure

├── main.py                     # FastAPI backend application
├── dashboard.py                # Streamlit frontend dashboard
├── start_server.py            # Server startup script
├── requirements.txt           # Python dependencies
├── requirements_python312.txt # Python 3.12 specific dependencies
├── env_example.txt            # Environment variables template
├── setup_gemini.py            # Google API setup helper
├── test_core_functionality.py # System testing script
├── src/
│   ├── models/
│   │   └── database.py        # SQLAlchemy models and database setup
│   ├── parsers/
│   │   ├── resume_parser.py   # Resume text extraction and parsing
│   │   └── jd_parser.py       # Job description parsing
│   ├── matching/
│   │   ├── hard_matching.py   # Keyword and exact matching
│   │   └── semantic_matching.py # Embedding and LLM-based matching
│   ├── scoring/
│   │   └── scoring_engine.py  # Final scoring and verdict generation
│   └── utils/
│       └── config.py          # Configuration management
├── data/
│   └── uploads/               # Temporary file storage
├── tests/                     # Test files
└── docs/                      # Documentation

🔌 API Endpoints

Core Endpoints

• POST /upload/job-description: Upload and parse job description
• POST /upload/resume: Upload and parse resume
• POST /evaluate: Evaluate resume against job description
• GET /results: Get evaluation results with filtering
• GET /dashboard/stats: Get dashboard statistics

Data Endpoints

• GET /job-descriptions: List all job descriptions
• GET /resumes: List all resumes
• GET /health: Health check endpoint

🧪 Testing

Run Core Functionality Tests

python test_core_functionality.py

This will test:

• API health check
• Job description upload and parsing
• Resume upload and parsing
• Resume evaluation workflow
• Dashboard endpoints

Manual Testing

1. Start the server: python start_server.py
2. Open dashboard: http://localhost:8501
3. Test both placement officer and student workflows
4. Verify real-time evaluation results

🚨 Troubleshooting

Common Issues

1. Port Already in Use

# Kill existing Python processes
taskkill /f /im python.exe

# Or change ports in start_server.py

2. Google API Key Issues

# Run setup helper
python setup_gemini.py

# Or manually add to .env file
GOOGLE_API_KEY=your_actual_api_key_here

3. spaCy Model Not Found

python -m spacy download en_core_web_sm

4. Python Version Compatibility

# Use Python 3.11+ or install specific requirements
pip install -r requirements_python312.txt

Getting Help

• Check API documentation: http://localhost:8000/docs
• Review error logs in terminal
• Ensure all dependencies are installed
• Verify Google API key is valid

📈 Performance

Benchmarks

• Resume Parsing: 1-3 seconds per file
• Job Description Parsing: 1-2 seconds per file
• Evaluation: 2-5 seconds per resume-JD pair
• Concurrent Users: Supports 50+ simultaneous users
• File Size Limit: Up to 10MB per file

Scalability

• Database: SQLite for development, PostgreSQL for production
• File Storage: Local storage with cleanup
• API: FastAPI with async support
• Frontend: Streamlit with caching

🔒 Security

• File Validation: Strict file type checking
• Input Sanitization: All inputs are validated and sanitized
• API Key Protection: Environment variable storage
• Data Privacy: Temporary file cleanup after processing
• Error Handling: No sensitive data in error messages

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

📄 License

This project is developed for Innomatics Research Labs. All rights reserved.

📞 Support

For technical support or questions:

• Check the troubleshooting section above
• Review API documentation at http://localhost:8000/docs
• Ensure all installation steps are completed correctly

---

🎉 Ready to automate resume evaluation at scale!

Start the system: python start_server.py
Access dashboard: http://localhost:8501