The problem
AI had to help with the learning task, not sit beside it as a demo. The app needed to stay reliable even when AI gave different answers.
What I did
- Website and app development
- AI features inside the learning process
- Frontend and backend development
- A code structure that supports future updates
Tools and systems
- AI features
- App state management
- Frontend and backend development
- App structure built for growth
Main technical choices
Give AI one clear job for the learner
Keep the app reliable when AI answers vary
Store app progress separately from the AI conversation
The result
AI became a useful part of the learning process. The app stored user progress separately from the AI conversation, so different AI answers could not break the rest of the experience.