Motivated by the large demand and limited supply of H-1B visas, I worked with classmates at Georgia Tech to build a predictive model and web application to assist prospective H-1B visa applicants. With a data set containing 3 million rows, we utilized logistic regression, support vector machines, and k-nearest neighbors to build an ensemble classification model which predicts the outcome of an H-1B visa application with 88.5% accuracy. These predictions feed into our web application, which enables prospective H-1B visa applicants to search for jobs by industry, job title, company, and location, thereby serving as a tool for identifying jobs that provide them with the best chance of obtaining an H-1B visa. The codebase and report are available via the GitHub link above.