The main objective of the project is to demonstrate the usefulness of AI/ML techniques in improving model forecasts at time and space scales, which are then applied to few selected area such as extremes, renewable energy and landslides. The application of AI/ML techniques has seen explosive growth in recent years including inimproving forecasts from physics based models. While many international forecasting centres (ECMWF& UK MetOffice) have made substantial investments in the use of AI/ML, the same is not the case with the modeling centres in India. This project proposes using state-of-the-art deep learning algorithms to improve and add value to the forecasts from existing models run by the MoES modeling centres.Additionally, a new generation of manpower versatile in both atmospheric and data sciences will be trained.