Open Day Presentation
BSc Freehand Drawing Tool with Reinforcement Learning for Classroom Teaching
Student
Desk
Initial task setting
Supervisor
Second Reader

g20-31

Keywords
AI whiteboard, drawing, education, SVG, learning

Technologies
Claude API, Firebase, Google Drive, JavaScript

Teaching with a digital whiteboard should feel as natural as teaching with a physical one. In practice, digital tools demand that teachers slow down, switch modes, and fight with interfaces not designed for a classroom. This project set out to change that by building a browser-based smart whiteboard where freehand drawing is the primary input and AI handles the rest.

Whatever the teacher draws or describes — an equation, a diagram, a labelled figure, a formal notation — the system interprets it and produces clean, professional output automatically, adjusted to the subject being taught. Subject-aware presets for Mathematics, Chemistry, Physics, Biology, and Computer Science bias the output towards the conventions of each discipline.

The system also learns from use. Instructions typed by users are anonymously pooled and periodically analysed by AI, which distils recurring preferences into style guidelines applied to every session.

The final implementation is a fully working browser application with Google Drive persistence, a configurable AI backend, and a conversation memory that allows teachers to refine and extend previous outputs interactively.