The FEAT project aims to develop efficient and sustainable fleet management strategies for shared electric micromobility systems (EmmS). The goal is to jointly optimize the energy consumption and the level of service of EmmS fleets, by taking into account dynamic energy grid loads, stochastic travel demands, battery energy waste in idle status, and coordination with other transportation modes. The approach is to use advanced machine learning models and state-of-the-art routing algorithms to empower the decision-making of where, when, and how to charge batteries, relocate vehicles, and swap batteries. The project is extended to 2025-11. Three papers benefited from the NAISS, where the ML models were trained with the cluster. The papers are under review now, and in the revision, we will need the cluster again to make the necessary changes under the same computational setups.