My research interests mainly include Machine Learning and Deep Learning techniques, and their applications in addressing real-world problems. More specifically, my PhD research has been focused on developing novel deep learning techniques for time series forecasting. The scope of the research was very wide with the focus on developing highly efficient recurrent neural networks that can be employed in Machine Learning Systems with online learning requirements, where we need to train the recurrent model frequently and ideally in the real-time. Moreover, such time series forecasting approach can effectively be employed for modeling nonlinear dynamical systems, where the state of the system is represented by highly nonlinear time series. The following slides portrays the general idea and the proposed approaches studied in this PhD research.