TeachingTurn

Inspiration

As a professor, I often struggle to get useful feedback while there is still time to improve a course. The end-of-semester evaluations come too late to help the students who gave them.

What it does

TeachingTurn lets students quickly share feedback after each lecture. It turns that feedback into one practical adjustment the professor can try in the next class. The plan is shared with students that can assess if it was effective.

How we built it

We built TeachingTurn as a Next.js web app. Students scan a QR code, answer two quick questions, and can add an optional comment. Supabase stores the responses, while OpenAI and GPT-5.6 helps identify supported themes and suggest a concrete teaching move.

I used Codex throughout all the dev pipeline, for planning, execution and deployment

Challenges we ran into

The hardest part was keeping the student experience fast and simple. Also, on how to close the loop between student feedback -> instructor action -> student outcome

Accomplishments that we're proud of

We created a feedback experience that takes only a few seconds. The tech stack and the production level quality that I could reach using Codex.

What we learned

I learned that professors do not need another complicated dashboard (using research with GPT 5.6). Technically, I learned about the full deployment pipeline of a functional app.

What's next for TeachingTurn

Next, I want to pilot TeachingTurn in real courses (specially my own courses), learn from students and instructors, and improve the quality of its recommendations.

Built With

• eslint
• nextjs
• openai
• pgtap
• playwright
• postgresql
• shadcn/ui
• supabase
• tailwind
• typescript
• vercel
• vitest