ninzizah/QuickGrade-Rwanda

# 🤖 AI-Powered MCQ System with Hugging Face An intelligent Multiple Choice Question system that uses **Hugging Face AI models** to automatically generate questions and evaluate answers. ## 🚀 Features ### 🤖 **AI-Powered Question Generation** - **Automatic MCQ Creation** using Hugging Face models - **Topic-based Generation** (Math, Science, History, etc.) - **Difficulty Levels** (Easy, Medium, Hard) - **Smart Question Suggestions** ### 📚 **Lecturer Dashboard** - Create papers manually or with AI assistance - Generate questions from 8+ subject areas - Real-time AI health monitoring - Export results and analytics ### 🎓 **Student Interface** - Take AI-generated MCQ tests - Timed examinations with auto-submit - Instant AI-powered feedback - Performance tracking ### 🛠️ **Admin Panel** - System-wide statistics - User management - AI service monitoring - Database analytics ## 🏗️ **System Architecture** ### **4-Layer Architecture:** 1. **🗄️ Data Layer** - MongoDB with Mongoose ODM 2. **🔌 API Layer** - Express.js REST API with AI integration 3. **🧠 Business Logic** - AI services and auto-grading 4. **🎨 Presentation** - React with Tailwind CSS ## 🤖 **AI Integration** ### **Hugging Face Models Used:** - **Question Generation**: `google/flan-t5-large` - **Answer Evaluation**: `microsoft/deberta-v3-large` - **Text Processing**: Various specialized models ### **AI Capabilities:** - Generate contextual MCQ questions - Evaluate open-ended answers - Provide intelligent feedback - Support multiple difficulty levels ## 🛠️ **Setup Instructions** ### **1. Prerequisites** ```bash - Node.js 16+ - MongoDB - Hugging Face API Key ``` ### **2. Environment Setup** ```bash # Clone and install npm install # Set up environment variables cp .env.example .env ``` ### **3. Configure .env** ```env # Database MONGODB_URI=mongodb://localhost:27017/mcq_system JWT_SECRET=your_jwt_secret_key # Hugging Face AI HUGGINGFACE_API_KEY=your_huggingface_api_key AI_QUESTION_MODEL=google/flan-t5-lar …