Salman’s research concentrates on integrating deep learning techniques into atmospheric sciences. He has developed a Deep Neural Network (DNN) solver that effectively solves complex PDEs, enabling accurate simulations of weather dynamics. Furthermore, he’s advancing the acceleration of CMAQ simulations using a digital twin approach, which enhances inverse modeling of emission inventory. Additionally, Salman is working on a real-time forecasting model for ozone pollution and has been actively involved in refining the bias correction techniques for the CMAQ model.