Tapestry
Tapestry is a daily briefing app. Each day, it scrapes news articles from a variety of different sources, and finds the most popular stories. It then generates unbiased AI summaries of the top-five stories of the day.
Cart-Pole Agent
Cart-Pole is a simulated Reinforcement Learning environment that Gymnasium offers. The goal is to move a cart back and forth, in order to balance a pole that has been placed on this cart. I built a Deep Q-Learning algorithm which successfully solves for the environment.
Image-Classifier of Trash
Using a dataset of images of trash, I buit and trained a convolutional neural network which classifies images by their type of trash, with an accuracy of 86%. This was in conjunction with an in-depth report looking into the issues and potential solutions of low Plastic Recycling.