Our team has spent the past two semesters working on machine modeling approaches to modeling COVID-19 infection rates in the United States. Our goal for this project was to create at least one model that performs better than the popular models currently available. We have gained professional experience working with data science and machine learning throughout the process. Specifically, we applied data collection, data preprocessing, analysis, machine learning model development, training, evaluation, and parameter tuning skills. We started with simple linear regression and autoregressive integrated moving average (ARIMA) models to set a baseline. Next, we shifted our focus to long short-term memory (LSTM) neural network models. Overall, we achieved two models that performed significantly better than existing models on the national scale. We then shifted to modeling COVID-19 infection rates in smaller regions, where we had more data to make the model more accurate.