YojnaSetu - AI for Bharat Hackathon

Bridging Citizens to Government Schemes through Multilingual AI

YojnaSetu is a voice-first, multilingual AI system that helps Indian citizens discover and access government schemes through natural conversation. Built for the "AI for Bharat" hackathon.

🎯 Problem Statement

Citizens across India struggle to discover government schemes due to:

• Language barriers preventing access to scheme information
• Complex eligibility criteria that are difficult to understand
• Lack of centralized discovery mechanisms
• Limited digital literacy in target demographics

🚀 Solution Overview

YojnaSetu uses conversational AI to:

• Understand natural speech in 5 Indian languages (Hindi, English, Tamil, Telugu, Bengali)
• Infer eligibility from conversational input without forms
• Handle uncertainty and incomplete information intelligently
• Provide personalized recommendations with clear explanations

🧠 Why AI is Essential

Traditional rule-based systems cannot handle:

• Natural language complexity: "My husband passed away last year, I have two children and work part-time cleaning houses"
• Contextual inference: Understanding that "my roof leaks during monsoon" implies housing scheme eligibility
• Multilingual processing: Code-switching like "Mera income bahut kam hai, around 15000 per month"
• Uncertainty handling: Making probabilistic inferences from incomplete information

🏗️ System Architecture

┌─────────────────┐    ┌──────────────────────┐    ┌─────────────────────┐
│   Voice/Text    │───▶│  Language & Speech   │───▶│  Eligibility        │
│   Interface     │    │  Intelligence        │    │  Inference Engine   │
└─────────────────┘    └──────────────────────┘    └─────────────────────┘
                                │                            │
                                ▼                            ▼
┌─────────────────┐    ┌──────────────────────┐    ┌─────────────────────┐
│   Response &    │◀───│  Scheme Knowledge    │◀───│  User Profile       │
│   Explanation   │    │  Base                │    │  Management         │
└─────────────────┘    └──────────────────────┘    └─────────────────────┘

📋 Key Features

• 🎤 Voice-First Interface: Hands-free operation for users with limited digital literacy
• 🌐 Multilingual Support: Hindi, English, Tamil, Telugu, Bengali with cultural context
• 🤖 Intelligent Matching: AI-powered eligibility inference beyond simple rules
• 📊 Confidence Scoring: Transparent uncertainty communication
• 📱 Progressive Web App: Works on any device with internet connectivity
• 🔒 Privacy-First: Minimal data collection with user consent

📁 Project Structure

.kiro/specs/yojna-setu/
├── requirements.md     # Detailed system requirements with acceptance criteria
├── design.md          # Technical architecture and AI model specifications  
└── tasks.md           # Implementation roadmap with 40+ actionable tasks

🛠️ Technology Stack

• Backend: Python, FastAPI, SQLite
• Frontend: React, TypeScript, Progressive Web App
• AI/ML:
  • Multilingual BERT for NLU
  • Web Speech API + Cloud Speech Services
  • Scikit-learn for eligibility classification
  • XGBoost for scheme ranking
• Deployment: Docker, Cloud hosting

📊 Success Metrics (Hackathon Demo)

• 70%+ accuracy in scheme eligibility matching
• 5-7 minute average interaction completion
• 80%+ speech recognition accuracy under demo conditions
• 50+ concurrent users support during demonstration
• 5 languages fully supported

🚀 Getting Started

1. Review the Specification

• Read requirements.md for detailed requirements
• Study design.md for technical architecture
• Check tasks.md for implementation roadmap

2. Development Setup

# Clone the repository
git clone https://github.com/YOUR_USERNAME/yojna-setu.git
cd yojna-setu

# Follow Task 1 in tasks.md to set up the development environment

3. Implementation Approach

The implementation follows a bottom-up approach:

1. Data Layer → Scheme database and models
2. AI Engine → Eligibility inference and language processing
3. User Interface → Voice-first web application
4. Integration → End-to-end system testing

🎯 Hackathon Goals

• Demonstrate AI Innovation: Show meaningful AI beyond simple chatbots
• Solve Real Problems: Address genuine barriers to government scheme access
• Scalable Architecture: Design for future expansion to production scale
• Social Impact: Enable better welfare distribution and citizen empowerment

🤝 Contributing

This project is developed for the "AI for Bharat" hackathon. The specification provides a complete roadmap for implementation.

📄 License

MIT License - Built for social impact and open collaboration.

---

Built with ❤️ for AI for Bharat Hackathon

Empowering citizens through intelligent technology