AI Personal Tutor
Multi-agent research system that creates personalized learning experiences using MCP orchestration, achieving real-time knowledge synthesis from diverse sources with interactive visualizations.
Overview
Engineered an intelligent tutoring system that transforms how people learn complex topics by orchestrating specialized AI agents through the Model Context Protocol (MCP). This project represents a synthesis of educational pedagogy and cutting-edge AI engineering, creating personalized learning paths that adapt to individual experience levels. The system demonstrates advanced agent coordination, where multiple specialized agents work in parallel to research topics, extract key concepts, build knowledge graphs, and generate tailored summaries. Each component is designed with educational principles in mind, ensuring learners receive structured, comprehensible content that builds upon their existing knowledge. Beyond its technical achievements, the AI Personal Tutor serves as a comprehensive example of production-ready agent orchestration, complete with real-time progress tracking, cloud storage integration, and a modern web interface that makes complex AI interactions accessible to non-technical users.
Technical Stack
Backend
- ▸Python 3.13
- ▸FastAPI
- ▸UV Package Manager
- ▸Pydantic
- ▸AsyncIO
AI/Agent System
- ▸Model Context Protocol
- ▸Multi-Agent Orchestration
- ▸Workflow Engine
- ▸WebSocket
- ▸JSON-RPC
Data & Visualization
- ▸Cytoscape.js
- ▸Knowledge Graphs
- ▸Markdown Generation
- ▸Google Drive API
- ▸Real-time Updates
Infrastructure
- ▸Docker
- ▸Redis Caching
- ▸PostgreSQL
- ▸Production Monitoring
- ▸API Documentation
Key Features
Multi-source research aggregation from web, YouTube, and academic content
Automatic knowledge graph generation with interactive visualizations
Experience-level adaptive summaries tailored to learner needs
Real-time progress tracking via WebSocket connections
Parallel agent execution for optimal performance
Google Drive integration for automatic backup and sharing
Quality filtering and source credibility ranking
Extensible architecture for adding new specialized agents
Code Examples
Technical Challenges
Coordinating multiple asynchronous agents with complex dependencies
Building reliable MCP communication across distributed services
Creating knowledge graphs that accurately represent concept relationships
Ensuring content quality while maintaining fast response times
Designing an intuitive interface for complex AI interactions