AGENTIC AI TUTOR

Architecture Diagram

Inspiration

The inspiration for my AI Tutor app comes from the need to bridge learning gaps among students through personalized education. Traditional teaching often fails to adapt to individual learning styles and paces. With AI, we can create an adaptive tutor that offers customized quizzes, feedback, and explanations. Our goal is to make learning more engaging, accessible, and effective for every student.

What it does

We built the AI Tutor using agentic AI architecture with multiple specialized agents — Tutor, Evaluator, Content, and Teacher agents. Each agent communicates through LangChain/LangGraph to perform adaptive learning, evaluation, and content generation. We used Amazon Q and Bedrock for intelligent quiz and content generation with real-time adaptability. The front end was built with Streamlit, and backend integration was done using Python and Azure/AWS services for scalability.

How we built it

We built the AI Tutor using a multi-agent architecture with Tutor, Evaluator, Content, and Teacher agents. Each agent collaborates through LangChain/LangGraph for reasoning, feedback, and adaptive learning. We integrated Amazon Q and Bedrock for dynamic quiz and content generation. The app runs on Streamlit frontend with Python backend deployed using Azure/AWS services for scalability.

Challenges we ran into

Integrating multiple AI agents and ensuring smooth coordination between them was complex. Generating truly adaptive and non-repetitive quizzes required fine-tuning LLM and Amazon Q outputs. Managing real-time personalization while maintaining speed and accuracy was challenging. Deploying the system on cloud with secure API and data handling took extra effort.

What we learned

We learned how to design and coordinate multiple AI agents for personalized learning. Gained deep insights into using Amazon Q and Bedrock for dynamic content generation. Understood the importance of prompt design, evaluation loops, and data feedback. Realized that building adaptive AI systems requires both technical precision and educational empathy.