MathTutorAgent

• 

  MathTutorAgent Web Interface

Inspiration

As a parent , I wanted to make math learning more personalized, engaging, and adaptive using Agentic AI and LLMs.

What it does

MathTutorAgent generates age-appropriate math questions, validates quality, classifies difficulty, detects duplicates, and stores learning history through a web-based AI tutor. 12

How I built it

I built it using Python, FastAPI, SQLite, and LLMs (OpenAI, Gemini, and Ollama) with a skill-based agent architecture and orchestration layer. 12

Challenges I ran into

I faced challenges with reliable skill routing, maintaining question quality, preventing duplicates, and supporting multiple LLM providers consistently.

Accomplishments that I'm proud of

I created a modular agentic tutoring platform with testing, tracing, persistence, validation, and a user-friendly web interface. 12

What I learned

I learned how to build scalable agentic AI systems, orchestrate specialized skills, and improve reliability through validation and observability.

What's next for MathTutorAgent

I plan to add step-by-step solutions, adaptive learning paths, progress tracking, gamification, voice tutoring, and support for additional STEM subjects.

Built With

• agentic-ai
• ai-agents
• edtech
• education
• fastapi
• generative-ai
• math-learning
• openai
• personalized-learning