AI Math Tutor

Inspiration

"Mathematics is often seen as a barrier for many students, not because the concepts are impossible, but because personalized help is hard to find. Many students find themselves stuck on a single problem at night with no one to ask for help. While search engines can give a quick answer, they rarely explain the 'why' behind the steps. Our inspiration came from the idea of 'Democratizing Education.' We wanted to use Artificial Intelligence to provide every student—regardless of their background or financial status—with a 24/7 personal tutor. We built AI Math Tutor to transform math from a subject of frustration into a journey of understanding, focusing on the logic of the solution rather than just the final result."

What it does

AI Math Tutor is an interactive web application where users can input complex math problems. The AI analyzes the equation and generates a detailed, step-by-step breakdown. It uses clear language to explain each mathematical rule being applied, helping students learn the process so they can solve similar problems on their own in the future."

How we built it

Challenges we ran into

One of the biggest challenges was 'Prompt Engineering'—ensuring the AI wouldn't just give a one-sentence answer. we had to fine-tune the instructions to make the AI act like a patient teacher who prioritizes explanation over speed."

Accomplishments that we're proud of

Integrating AI responses into a strictly typed TypeScript (TSX) environment was a challenge. We had to define custom interfaces for the AI's JSON output to ensure that the data flowing through our app was secure and bug-free. This helped us catch errors early and made the codebase much more robust."

What we learned

"Building AI Math Tutor in such a short timeframe was an intense but rewarding learning experience. Here are the key takeaways from our team: Type Safety with TSX: We learned the importance of using TypeScript (TSX) in a fast-paced environment. Defining strict types for our components and API responses saved us hours of debugging that would have otherwise occurred with plain JavaScript. Prompt Engineering: We discovered that working with AI is more than just sending a request. We learned how to 'engineer' prompts to ensure the AI behaves as a tutor (explaining steps) rather than just a calculator (giving answers). Mathematical Rendering: We learned how to handle complex string data from AI and render it into beautiful, readable mathematical formulas using LaTeX-style formatting. Time Management: With only 11 hours on the clock, we learned how to prioritize a Minimum Viable Product (MVP)—focusing on core tutoring functionality before adding extra aesthetic features. API Integration: This project gave us hands-on experience in connecting a modern React frontend with powerful AI backend services, managing asynchronous data, and handling loading states for a smooth user experience

What's next for AI Math Tutor

Handwritten Problem Recognition (OCR) The Goal: Allow students to take a photo of their notebook instead of typing. Tech: Integrate Google Vision API or OpenAI’s GPT-4o-mini to convert handwriting into digital math instantly. 2. Interactive "Socratic" Mode The Goal: Instead of giving the full answer immediately, the AI should ask, "What do you think the first step is?" Impact: This transforms the app from a "calculator" into a true teacher that encourages active thinking. 3. Voice-Guided Explanations The Goal: Let the AI "talk" to the student. Tech: Use Web Speech API or ElevenLabs to provide audio explanations for students who learn better by listening or for those with visual impairments (Track 1: Social Good). 4. Gamified Progress Tracking The Goal: Add badges, "Daily Streaks," and experience points (XP). Impact: High school and middle school students are much more likely to keep practicing math if it feels like a game. 5. Multi-Subject Expansion The Goal: Use the same architecture to build an AI Physics Tutor or AI Chemistry Tutor. Why: Since you’ve already built the "Tutor Logic" in TSX, switching the subject is just a matter of changing the AI's prompt!