Mosaic Learn

What it does

I am a rising senior in high school (12th grade) interested in computer science and software engineering. While I am a self-directed learner, my classmates often ask my help to explain concepts in a more "user-friendly" manner, leading me to harness the power of "teaching them to fish"–guiding them to understand the underlying principles, instead of explaining individual problems. To address this educational gap, I worked on a mobile application that allows users to scan textbook pages and receive links to YouTube videos for further learning. This way, my peers can more easily learn core concepts without being completely lost.
Mosaic Learn is an application built on top of React Native and Expo for the Front end and Firebase, Gemini Dev API and YouTube Search API on the backend. When a user takes a picture of a document in our UI, we use Google MLKit for cross-platform text extraction. We then send this information to the Gemini 1.5 Flash API for parsing and return a JSON with a summary of the document and an array of search terms. We then feed these search terms into the YouTube results API and save the returned results to Firebase, so users can access their scan history. Mosaic Learn is my first step into using GenAI to help students like me learn concepts more quickly and thoroughly. We hope that Mosaic Learn will make education more accessible, engaging, and effective for students worldwide.