My name is Matthew (Matt) McCune, I am a graduate from the University of Rhode Island with a Bachelors of Science in Computer Science. I have a passionate interest in deep learning, software development, and designing solutions to complex problems. In my free time I like to surf, ski, be outside, and go to the gym.

ConvTherm is a Convolutional LSTM deep learning model trained to predict temperature across the continuous United States, this model was designed to be a lightweight and lightning fast alternative to traditional physics based forecasting models. The webpage displays model inference showing precomputed model forecasts, observed temperatures, and the difference between them for different initialization dates. This model is meant as a proof of concept that with more channels, global scope, and training years, a extremely accurate model capable of producing more output channels than just temperature is possible.