Abstract:
Basha Buddy is a gamified mobile application designed to improve English language learning
for Sinhala-speaking students. Recognising the challenges faced by young learners in
multilingual environments such as Sri Lanka, the app combines interactive learning with
playful engagement. It offers level-based tasks, voice prompts in Sinhala and English and real-
time progress monitoring to maintain motivation and encourage learning.
To allow students to practice writing directly on the screen, a custom Convolutional Neural
Network (CNN) based on the LeNet architecture has been implemented to recognize
handwritten English capital letters. Firebase was used for cloud storage and authentication,
FastAPI for back-end services and model inference, and Flutter for the mobile front-end of the
application. Gamification features like leaderboards, scoring systems, and rewards are meant
to keep students interested, and a special admin panel lets teachers assign tasks, track progress,
and control content.
By integrating handwriting recognition, language support and game-based learning
mechanisms, Basha Buddy offers a novel, locally relevant solution that promotes active
language learning in a child-friendly and technologically advanced way. The project not only
fulfils pedagogical goals, but also represents an innovative model for combining artificial
intelligence with interactive learning tools in primary education.
