This project builds a smart trash can that automates the process of separating trash into compost and recyclables, thus obviating the need for multiple bins and reducing the burden of sorting on the customer. The smart trash can also helps limit the amount of contamination (unsorted waste) in waste streams, which is one of the primary difficulties recycling facilities face given the high costs of sorting equipment and manual labor required. The project team develops the algorithms, specifically via Vanilla Convolutional Neural Networks, needed for the trash can to classify images of trash. The smart trash can receives a piece of trash, identifies it as plastic or compost after an image is taken, then tilts either left or right to place the object in the correct bin underneath. In doing so, the project contributes to building a dataset of images of trash that can be used in waste statistics and future projects.