Deep learning has emerged as a powerful artificial intelligence tool that can be applied to various types of railroad maintenance and operations issues. Several deep learning models have recently emerged, and these models have very powerful text mining capabilities that include classification problems using text as input, and the ability to automatically answer questions from a text. There are considerable amounts of text data available like maintenance reports, accident records and operational reports that can be mined to gain new insights. This article describes the application of two types of deep learning models to predict maintenance effects on operations and to automatically answer questions about maintenance standards. This webinar will introduce deep learning and text mining. Several different railroad applications of deep learning including Long-Short Memory networks and transformer models will be discussed. Examples include the prediction of operational effects of maintenance activities, automated question answering from design manuals.