As we took our engineering classes, we noticed most of out colleagues using LLMs like ChatGPT write their homework for them. We don't believe that this is either an effective, nor an honest way of studying.

So we made StudyMate. StudyMate is a learning tool that reads a student's handwriting, and pinpoints their mistakes. It allows a student to work on a problem themselves-- allowing real learning, while still integrating modern technology. Only after the student completes their work does StudyMate analyze their work and highlight their mistakes.

StudyMate is built with a React + TypeScript frontend and a FastAPI backend. The frontend provides a digital canvas where students can write equations naturally by hand. Handwritten math is sent to the backend and processed using Mathpix, which converts handwriting into structured LaTeX. The backend analyzes each algebraic step using custom rule-based logic, and is capable of both differentiating, and solving equations. Instead of solving the problem outright, StudyMate detects incorrect transitions between steps and identifies the type of mistake made.

The biggest challenges today was managing time constraints while also teaching ourselves how to create a fully functional product. As first year students, this was rather difficult. another Challenge we ran into was handwriting variability, as people write out equations in many different ways. We needed one program that can read everything.

We are very proud to deliver a fully functional MVP for our very first hackathon as first year students.

We learned that detecting how a student is wrong is often far more important than knowing that they are wrong.

What is next for StudyMate is to support more advanced math topics, like differential equations or multi-variable calculus. We would also like to further integrate AI models in order to provide a more detailed tutoring experience

Built With

Share this project:

Updates