EducAI - AI Educational Assistant
EducAI is a comprehensive web application for educational assistance based on AI, specifically designed for generating and correcting mathematics exercises for Cycle 2 students (CP, CE1, CE2). The platform leverages OpenAI's GPT models (GPT-3.5-turbo, GPT-4o) through LangChain to provide intelligent exercise generation, automated correction of student submissions, personalized recommendations, and curriculum management. The system uses RAG (Retrieval-Augmented Generation) with FAISS vector store for curriculum-based exercise generation, OCR for processing scanned student work, and a full-stack architecture with FastAPI backend and React frontend.
Overview
EducAI is a comprehensive web application for educational assistance based on AI, specifically designed for generating and correcting mathematics exercises for Cycle 2 students (CP, CE1, CE2). The platform leverages OpenAI's GPT models (GPT-3.5-turbo, GPT-4o) through LangChain to provide intelligent exercise generation, automated correction of student submissions, personalized recommendations, and curriculum management. The system uses RAG (Retrieval-Augmented Generation) with FAISS vector store for curriculum-based exercise generation, OCR for processing scanned student work, and a full-stack architecture with FastAPI backend and React frontend.
Key Features
AI-powered exercise generation (QCM) with natural language interface
RAG integration with FAISS vector store for curriculum-aware generation
Automated correction of PDF submissions using OCR and LLM analysis
Student performance tracking and analytics dashboard
Personalized AI-generated recommendations for students
Curriculum management with PDF upload and vector search
JWT-based authentication and role-based access control
Chat interface for interactive exercise generation
Multi-agent system using LangChain and LangGraph
Structured output parsing with Pydantic
Real-time exercise generation with curriculum context
Comprehensive database models for teachers, students, exercises, and results
pages.portfolio.projects.educai_ai_educational_assistant.features.12
Technical Highlights
Developed comprehensive AI-powered educational platform with OpenAI GPT integration
Implemented RAG system with FAISS vector store for curriculum-aware exercise generation
Created automated correction system using OCR and LLM-based structured extraction
Built full-stack application with FastAPI backend and React frontend
Integrated multi-agent system using LangChain and LangGraph
Designed comprehensive database schema for educational workflows
Challenges and Solutions
LLM Output Parsing
Implemented Pydantic output parsers and JSON extraction with regex for consistent LLM response parsing
Curriculum Integration
Developed RAG system with FAISS vector store, PDF processing, and semantic search for curriculum-aware generation
PDF Processing
Integrated OCR with Tesseract and LLM-based structured extraction for automated student submission processing
Authentication & Authorization
Implemented JWT tokens, password hashing, and role-based access control for secure platform access
Real-time Exercise Generation
Created LangChain agents with RAG retrieval and prompt engineering for on-demand exercise generation
Student Performance Analysis
Developed LLM-based analysis with structured output parsing and performance aggregation for recommendations
Technologies
Backend
AI/ML
Document Processing
Frontend
Deployment
Environment
Project Information
- Status
- Completed
- Year
- 2024
- Architecture
- Full-Stack Microservices Architecture with FastAPI Backend, React Frontend, and AI Integration
- Category
- Full-Stack Development