Electrolytes play a crucial role in the development of electrochemical energy storage systems such as batteries and supercapacitors. Apart from desired bulk properties of the electrolyte such as a good ionic conductivity, a high electrochemical (thermal) stability and electrolyte interfaces are of paramount importance in electrochemical energy storage systems In order to achieve these technological requirements and to design better electrolyte materials, a molecular-level understanding of the transport mechanism and the electrochemical stability is necessary. In principle, density functional theory (DFT) calculations and DFT based molecular dynamics (MD) simulations are suitable techniques for this task, where the distinction between reactive solute and solvent has all but disappeared. However, the associated computational cost is too high to compute the fully converged ionic conductivity or explicitly simulate the reactive electrode/electrolyte interface. This calls for atomistic machine learning, which can provide quantum mechanical accuracy without electrons and propose a broad range of likely candidates beyond human intuition, because of the automated feature engineering and the universal approximation ability. In this project and our continuous efforts, we will combine physics-based simulation (reference DFT/DFTMD simulations) and machine learning-based simulations (Tensorflow and Numpy) to model electrolyte materials. Our code development and application will be based on a Python library for building atomistic neural networks of molecules and materials (PiNN, https://github.com/Teoroo-CMC/PiNN/). The project is intended for 12 months and the team consists of Master, PhD students, and a Principle Investigator (PI). This young team is supported by a recent starting grant from the Swedish Research Council (VR). Our recent works in this area can be found as follows: [1] Thomas Dufils, Lisanne Knijff, Yunqi Shao, and Chao Zhang*, "PiNNwall: Heterogenous Electrode Models from Integrating Machine Learning and Atomistic Simulation", J. Chem. Theory Comput. 2023, 19: 5199 [2] Yunqi Shao, Linnéa Andersson, Lisanne Knijff, Chao Zhang*, "Finite-field coupling via learning the charge response kernel", Electro. Struct., 2022, 4: 014012 (Invited article) [3] Lisanne Knijff, Chao Zhang*, "Machine learning inference of molecular dipole moment in liquid water", Mach. Learn.: Sci. Technol. 2021 2: 03LT03 [4] Yunqi Shao, Florian M Dietrich, Carl Nettelblad, Chao Zhang*, "Training algorithm matters for the performance of neural network potential: A case study of Adam and the Kalman filter optimizers", J. Chem. Phys., 2021, 155: 204108 [5] Yunqi Shao, Lisanne Knijff, Florian M. Dietrich, Kersti Hermasson, Chao Zhang*, “Modelling Bulk Electrolytes and Electrolyte Interfaces with Atomistic Machine Learning”, Batter. & Supercaps, 2021, 4: 585 (Minireview) [6] Yunqi Shao, Matti Hellström*, Pavlin D. Mitev, Lisanne Knijff, Chao Zhang*, “PiNN: A Python Library for Building Atomic Neural Networks of Molecules and Materials”, J. Chem. Inf. Model., 2020, 60: 1184