Time series data consists of sampled data points taken over time. Time series modeling and prediction are one of the most fundamental problems and have been subject to intensive researches. In recent decades the real time world time series data have become increasingly complex, high dimensional, noisy and high throughput. The modeling and making sense of modern time series are still very challenging. In recent years deep learning has been found huge successes in many applications such as image processing, computer vision, facial recognition, etc. Many deep learning paradigms are of high number of layered artificial neural networks, and the learning is based on hybrid learning approaches such as pre-training, semi-supervised, which allow representation learning to be possible from very complex data sets. This project will focus on applying existing deep learning models as well as developing new deep learning algorithms for time series data sets in particular. The research scope is large, so there is flexibility in the project depending on the student interest. For example, research topics that may be of interests include adaptive nonstationary time series prediction, time series anomaly detection and unsupervised learning techniques for high dimensionality reduction, and multistep ahead prediction etc. These researches will provide solid footing for highly demanding data scientist jobs. Applicants with applied mathematics and/or electrical engineering degrees are particularly welcome.