We built a platform for teachers at a hackathon. We didn't win
Four days to figure out how AI can help a teacher with a full classroom. I built the live part: QR check-in, the assistant and the teacher's live view.
In June, at a school hackathon, we built TICApp as a team of four. We started from something we see every day: there are more and more students per teacher, and it's very hard to know in the moment who understood, who needs help and who's ready to move on. Without that information, you teach to the average.
Three parts
- Insight. The teacher uploads their material —a PDF, slides or text— and the AI creates comprehension questions. Students join with a QR code and answer from their phones, and the teacher sees in real time which concepts students are struggling with.
- Guardian. During a digital exam it flags when a student switches tabs, copies and pastes or leaves full screen. It tells the teacher and doesn't block anyone: a person makes the call.
- Tito. An assistant for students' questions. During an exam it never gives away the answer.
My part
I handled what happens live during class: students joining by scanning a QR code, Tito answering with real AI instead of canned replies, and the screen where the teacher sees the classroom in real time.
We didn't win. The code stayed where it was, and we didn't keep working on it.
What I take from it
- AI in the classroom should give the teacher information, not replace them. That's why Guardian flags instead of punishing.
- Making a demo actually work, with real AI and live data, takes more effort than showing a staged one. I'd rather do that.
All notes
All notes →