MathMint – Hackathon Project
Inspiration
We wanted to make mathematics more meaningful for Class 10 students by showing how it applies in everyday life. Banking and finance felt like the perfect example—every student has seen parents talk about savings, loans, or fixed deposits, but rarely connects it to formulas in their textbook. The idea of using AI to turn simple maths problems into interactive learning experiences inspired us to create MathMint.
What it does
MathMint is an AI-powered platform that helps students understand Simple Interest vs Compound Interest in a fun, visual, and practical way.
- Takes real-world financial scenarios (e.g., “₹10,000 for 5 years at 8% p.a.”).
- Explains the step-by-step math at a Class-10 level.
- Checks student’s own working line by line and gives hints.
- Generates graphs comparing SI vs CI.
- Exports the whole exercise as a neat whitepaper using a ready template, so students can showcase their project work.
Example formulas:
Simple Interest (SI):
[ SI = \frac{P \times R \times T}{100} ]Compound Interest (CI):
[ A = P \times \left(1 + \frac{R}{100}\right)^T, \quad CI = A - P ]
How we built it
- Hybrid approach: deterministic math calculations for accuracy, AI (LLM) for explanations and error feedback.
- Frontend: React + Tailwind for a clean, one-page interface.
- Backend: Python FastAPI with endpoints for solving, checking steps, and exporting reports.
- AI layer: prompt-engineered to parse natural-language scenarios, generate CBSE-friendly explanations, and validate student steps.
- Export:
python-docxto auto-fill a formatted project report (whitepaper) and convert to PDF. - Graphs: Chart.js/Plotly to render SI vs CI growth over time.
Challenges we ran into
- Preventing AI from “hallucinating” numbers — solved by letting deterministic math handle all numeric outputs.
- Designing a step-checker that gives useful hints without revealing the full answer.
- Time management between UI polish, backend accuracy, and AI integration in a short hackathon.
- Exporting graphs and explanations neatly into Word/PDF format.
Accomplishments that we're proud of
- Built a working MVP that’s educational and practical.
- Bridged CBSE curriculum with real-world financial literacy.
- Exported a polished whitepaper template—turning math practice into something students can showcase.
- Created a solution that teachers can adopt in classrooms immediately.
What we learned
- How to integrate AI with rule-based systems for accuracy and explainability.
- The importance of pedagogical framing—AI outputs must be tailored to grade level.
- Rapid prototyping: turning an idea into a student-friendly product within days.
- Small differences in prompt design can affect clarity and correctness of AI explanations.
What's next for MathMint
- Expand beyond SI/CI: add EMI calculators, GST breakdowns, savings goals, and SIP growth.
- Add multilingual support (Hindi + regional languages).
- Introduce OCR for handwritten homework (snap a photo → instant feedback).
- Build a teacher dashboard for class tracking and booklet exports.
- Partner with schools, CBSE exhibitions, and children’s science congresses to bring MathMint to more students.
Log in or sign up for Devpost to join the conversation.