Sahil Jadhav: AI-Powered Digital Twin (RAG) 🤖💼

A production-grade Retrieval-Augmented Generation (RAG) system acting as a professional "Digital Twin."

---

🌟 Overview

This project is a high-performance RAG application designed to represent me professionally. It uses a hybrid search architecture to provide 100% accurate, source-cited information from my resume, ensuring recruiters and interviewers get the data they need instantly via Web and Mobile.

🎥 Key Features

• Hybrid Search: Combines semantic vector search (BGE-Small) with keyword matching for zero-miss retrieval.
• Contextual Chunking: Breaks down resume sections into atomic, context-enriched pieces.
• Real-time Streaming: Chat responses stream token-by-token for a "ChatGPT-like" experience.
• Cross-Platform: Modern Streamlit Web Dashboard + Expo Mobile App.

---

🛠️ Tech Stack

• LLM Engine: Llama 3.2 (via Ollama)
• Vector Database: Supabase (pgvector)
• Backend API: Flask (Production-hardened with Gunicorn & Gevent)
• Frontend: Streamlit (Web) & Expo/React Native (Mobile)
• Infrastucture: Docker & Docker Compose

---

🚀 Quick Start (Development)

1. Prerequisites

• Ollama installed and running (ollama pull llama3.2)
• Python 3.10+
• Node.js & npm (for Mobile)

2. Setup

# Clone the repository
git clone https://github.com/yourusername/resume-rag.git
cd resume-rag

# Install Python dependencies
pip install -r requirements.txt

# Start the Flask Backend
python -m app.api

3. Mobile App (Expo)

cd mobile
npm install
npx expo start --tunnel

Note: Run python update_tunnel.py in the root folder whenever your ngrok URL changes.

---

🏗️ Production Deployment (Docker)

To run the entire stack in a productionalized environment:

docker-compose up --build -d

Service	Endpoint
Backend API	http://localhost:5000
Streamlit UI	http://localhost:8501
Health Check	http://localhost:5000/health

---

📁 Project Structure

resume_rag/
├── app/                  # 🧠 Backend & RAG Logic
│   ├── api.py            # Flask API (Streaming capable)
│   ├── config.py         # Centralized environment config
│   ├── rag_answer.py     # Prompt engineering & LLM bridge
│   └── streamlit_app.py  # Web Interface
├── mobile/               # 📱 Expo Mobile App
│   ├── app/(tabs)/       # Chat UI Screens
│   └── services/api.ts   # Mobile-to-Backend service
├── data/                 # 📄 Source Artifacts (resume.md)
├── Dockerfile            # Production API image
├── Dockerfile.ui         # Production Streamlit image
└── update_tunnel.py      # Automated DX tool for mobile tunnels

---

📝 Configuration

Create a .env file in the root:

DATABASE_URL=your_supabase_postgresql_url
APP_ENV=dev
CORS_ORIGINS=*

---

🙌 Author

Sahil Jadhav
AI/ML Engineer