Short-term forecasting of clouds and other hazardous weather conditions is important to U.S. Navy operations.  This position will involve developing machine learning (ML) based weather models to operate on different combinations of available datasets.  ML weather models provide a computationally lightweight alternative to operational numerical weather prediction (NWP) systems, and can be trained on many different sets of observations.  The successful applicant will develop an ML weather system that uses various inputs to generate the best possible weather forecast guidance.  The project aims to quantify the sensitivity of an ML weather model's performance based on the observations available in the training dataset.