A Next-Gen ATS & Career Accelerator built with Hybrid AI.
- Hybrid Scoring Algorithm: Combines Keyword Match (50%) with Semantic Vector Similarity (50%) to grade resumes fairly.
- Role Inference: Automatically detects if a candidate is a Data Scientist, Product Manager, etc., using Llama-3 logic.
- If a resume scores > 85/100, the system automatically adds it to its Memory (RAG).
- It extracts new technical skills from high-quality resumes to grow its internal
skills_db.json.
- Global Leaderboard: Ranks candidates against the entire database of applicants, providing a true percentile score.
- Batch Processing: Upload 50+ resumes at once; the system synchronizes rankings so every candidate is compared against the full batch.
- Professional Templates: Generates clean, ATS-friendly PDF resumes using standard LaTeX-style formatting.
- AI Content Generation: Uses RAG to suggest "High-Impact Bullet Points" based on community best practices.
- Smart Fallback: If PDF generation fails (due to driver issues), it automatically provides a high-fidelity HTML download.
- Cover Letter Generator: Writes personalized cover letters analyzing the gap between your resume and the job description.
- Interview Prep: Generates 5 tailored technical & behavioral interview questions based on your specific weak points.
- Frontend: React (Vite), Tailwind CSS, Framer Motion, Lucide Icons. Theme: Indigo Glass (Premium UI).
- Backend: FastAPI, Uvicorn, Python 3.10+.
- AI Engine:
- LLM: Groq (Llama-3-8b-8192) or Google Gemini 1.5 Pro.
- Embeddings:
all-MiniLM-L6-v2(SentenceTransformers). - Vector DB: FAISS (Facebook AI Similarity Search).
- Database: MySQL (for persistence), JSON (for distinct rules).
- Python 3.10+ installed.
- Node.js 16+ installed.
- MySQL Server running locally (e.g., via MySQL Workbench or XAMPP).
# Clone the repository
git clone <repo-url>
cd resume-screening
# Create Virtual Environment
python -m venv venv
# Activate (Windows)
venv\Scripts\activate
# Activate (Mac/Linux)
source venv/bin/activate
# Install Dependencies
pip install -r requirements.txtCreate a .env file in the root directory:
GROQ_API_KEY=gsk_your_key_here
# Database Config
DB_HOST=127.0.0.1
DB_PORT=3306
DB_USER=root
DB_PASS=your_password
DB_NAME=resume_dbcd frontend
npm install# From the root folder
uvicorn main:app --reload- Server will start at
http://localhost:8000. - Swagger API Docs available at
http://localhost:8000/docs.
# From the frontend folder
npm run dev- App will launch at
http://localhost:5173.
This system is not static. It gets smarter with every "Perfect" resume it sees.
- Ingestion: When a user scans a resume and it receives a Score > 85:
- Vector Memory: The full text is embedded and saved to
data/memory.index. Future users generating content will get advice inspired by this resume. - Rule Expansion: The LLM extracts distinct tools/skills (e.g., "Polars", "Qdrant") and adds them to
data/skills_db.jsonif they don't exist. - Verification: You will see a purple "⚡ AI LEARNED FROM THIS" badge on the result card.
├── data/ # storage for vector index & skills DB
├── engine/ # LLM & Vector Logic
├── frontend/ # React Application
├── routers/ # API Endpoints (Audit, Builder, Companion)
├── templates/ # HTML Templates for PDF Generation
├── utils/ # Helper scripts (PDF extraction, DB storage)
├── main.py # FastAPI Entrybolint
└── requirements.txt # Python Dependencies
Made By Pranay Bagaria
MIT License. Created for the Advanced Resume Intelligence Project.