Project

Overview

This project was created as a final project for my 9th grade AP CSA class, and my friend Laveden partnered up with me. It's a social media in the terminal. This project uses a simple enough backend consisting of mongodb and Express.js, but there is one part that is a bit more complex: the post recommendation system. When each post is created, an embedding model is used to store (alongside the English version) a numerical representation of that post. Whenever a user likes a post, its vector is incorporated into their interest vector through element-wise operations. (It is also removed if they unlike a post.) Then, when a user goes to the feed, the posts are sorted by the cosine similarity between the user's interest vector and the post's vector. Overall, this project was a good practice opportunity for my existing backend development skills but an amazing learning opportunity to delve deep into the fascinating world of machine learning (specifically vectors and TNNs) and how they can be used to create a better user experience. I was responsible for all backend services, while my friend made the frontend.