Learning-Aware Generative AI System

Overview

Learning-Aware Generative AI System is a 3-layer AI architecture built for learning content generation. It features automated preprocessing, a RAG pipeline for contextual prompt enrichment, and a secondary LLM layer for hallucination detection and correction. The system engineers end-to-end pipelines using vectorized embeddings and automated LLM orchestration to enable high-throughput content generation.

Key Features

• 3-Layer AI Architecture:
  • Preprocessing & Chunking: intelligently breaks down curriculum content (PDFs).
  • RAG Pipeline: Contextual prompt enrichment using Vector DB (ChromaDB).
  • Hallucination Detection: Secondary LLM layer to verify and correct generated outputs.
• Automated Content Generation:
  • Video Lessons: Generates synchronized math explanatory videos using Manim and Edge-TTS.
  • Quizzes & Flashcards: Automatically generates practice quizzes and revision flashcards from curriculum content.
  • Teaching Plans: Creates structured teaching plans for educators.
• Interactive Chatbot: Context-aware RAG chatbot ("MathBuddy") for student Q&A, strictly grounded in the official curriculum.
• High Performance: Engineered for high-throughput, low-latency generation.

Technology Stack

• Languages: Python
• Backend Framework: FastAPI (High-performance web framework) / Flask
• AI/ML:
  • LLMs: Llama-3 (via Groq API) for reasoning and generation.
  • RAG: ChromaDB (Vector Store), Sentence Transformers (Embeddings).
• Data Procesing: BeautifulSoup, Selenium (Web Scraping), PDFPlumber (PDF Extraction), Pandas, NumPy.
• Media Generation: Manim (Mathematical Animation Engine), Edge-TTS, MoviePy.

Installation

1. Clone the repository:

   git clone <repository_url>
   cd Aletheia

2. Install dependencies: Ensure you have Python 3.10+ installed.

   pip install -r requirements.txt

   Note: Manim requires FFmpeg and LaTeX to be installed on your system.

3. Environment Setup: Create a .env file in the root directory and add your API keys:

   GROQ_API_KEY=your_groq_api_key
   YOUTUBE_API_KEY=your_youtube_api_key

Usage

Starting the Application

The application uses a FastAPI backend. To start the server:

# Run the backend server
python backend/main.py

Or potentially use the provided batch script if on Windows:

start_app.bat

The API will be available at http://localhost:8000 (Documentation at http://localhost:8000/docs).

Running Individual Modules

You can also run specific pipelines directly:

• Chatbot (CLI Mode):

  python chatbot_rag.py

• Video Generation:

  python generate_animations_synchronized.py

Project Structure

• backend/: FastAPI server and API endpoints.
• content/: Stores curriculum source files (PDFs) and processed JSON.
• generated_content/: Output directory for generated Videos, PDFs, and Plans.
• chatbot_rag.py: RAG implementation for the interactive chatbot.
• extract_pipeline.py: Core logic for extracting and structuring content from PDFs.
• manim_engine_synchronized.py: Logic for driving Manim animations based on generated scripts.

---

Built with ❤️ for Education.