In this thesis, I present a portfolio of projects centered around the theme of Machine Learning for Social Good (ML4SG). I present two fully functional prototypes for machine-learning based systems targeting societal challenges. The first project is Art I Don’t Like, an anti-recommender system for visual art that I built in Fall 2018 and Winter 2019. This anti-recommender system encourages users to reflect on their relationship to personalization algorithms on the internet through their art preferences. The second project is CompostNet, an image classification model for personal meal waste. I developed CompostNet in Spring 2019. CompostNet is a mobile application that includes an image classifier for meal waste, to help individuals divert their personal waste from landfills. In the introduction I discuss some categories ML4SG projects share - method, intended users, and phase of deployment. I reflect on my projects to determine that Art I Don’t Like is a Social good-first, individual-oriented, project at small-scale deployment, while CompostNet is a Social Good-First, Individual-oriented, undeployed project.