HeyArya!
I built an AI math tutor that explains things the way Feynman would: simply, honestly, and at whatever level you are actually at. No answer dumping. Just learning.
Why I built it
Saksham and Live the Learning taught me that access to learning is uneven. HeyArya! was my attempt to scale the other side: a tutor that anyone with a browser could reach, at midnight before an exam, in a village with no coaching center, or in a classroom where the teacher has forty students and ten minutes.
I did not want another app that solves problems for you. I wanted something that walks you through the thinking, adjusts when you are lost, and treats confusion as the starting point rather than a failure.
What it does
HeyArya! is an AI-driven educational platform for interactive math tutoring. I designed it around three ideas pulled straight from how Feynman taught:
- Clear explanations. Complex concepts broken down without dumbing them down.
- Interactive learning. Conversation, not a wall of generated text.
- Personalized tutoring. Responses tuned to your level, not a generic syllabus script.
Students can jump into calculus, algebra, geometry, or statistics and ask what they actually need: a concept explained, a step unpacked, a problem walked through from first principles.
What I built
I built and shipped the full stack myself: the tutoring interface, the adaptive prompting layer, and the subject routing that sends a question to the right kind of explanation. I tuned the system so it nudges toward understanding rather than handing over a finished solution, which matters when ChatGPT shortcuts are the default.
The Feynman thread is not cosmetic. It shapes how the model is instructed to respond: use plain language, check whether the student followed, and re-explain from a different angle when they did not.
Where it stands
HeyArya! has passed 9,000 views with around 100 active users across 9 countries. It is live, free to try, and still the project I iterate on when I think about who gets a tutor and who does not.