The process of maintenance is always considered to be a huge driver of costs in all industries. Depending on the industry, maintenance activities can account for 15–70% of the total production costs. Despite that, most of the industries still rely upon maintenance policies that are outdated and severely inefficient from a time and money point of view. In this context, the railway industry is no exception. Maintenance of high-value assets of Indian Railways is still done primarily through conventional maintenance practices. This causes the production time to go down and the overall quality of the components to deteriorate. On the other hand, there is ample research work being done to explore the details of several other maintenance policies. One of the most efficient and highly preferred maintenance policies is predictive maintenance. This study reviews existing literature on predictive maintenance and its implementation in the railway industry and identifies gaps and prospects for further research. The objective of this study is to begin with understanding the current maintenance policies used by Indian Railways, and then go about outlining the potential advantages of implementing predictive maintenance. To signify the importance of predictive maintenance, an analysis is performed over real-world data of rolling stock by training a machine learning model over the data and predicting the Remaining Useful Life of the components. The model is trained using a type of Recurrent Neural Network, known as Long Short-Term Memory networks. This training is carried out by a regression algorithm. Finally, the predictions from the model are plotted and compared with the actual data, to indicate the efficacy of the model. After interpreting the findings of the plot, it is concluded that such predictive maintenance systems could be installed in the rolling stock operated by the Indian Railways, as it would impact the overall availability and efficiency of the assets and boost the operations of the organization.