Understanding and predicting the behaviour of functional materials, interfaces, and energy systems requires accurate descriptions across multiple length and time scales. Despite major advances in electronic structure theory, molecular simulation, and machine learning, significant challenges remain in bridging these scales and capturing complex environments such as interfaces, disordered materials, and reactive conditions. In particular, the treatment of charge transfer, spin states, and interfacial effects are still open problem. This proposal brings together three interconnected research areas in materials chemistry, unified by a strong methodological focus. A central component is the use of foundation models—machine-learning interatomic potentials trained on large datasets—which can be adapted to new systems through fine-tuning using a limited number of additional DFT calculations. While these models offer great potential, there is currently no established best practice for efficient fine-tuning, especially for chemically complex systems involving redox processes and interfaces. A significant part of the project will therefore focus on developing and validating strategies for fine-tuning and applying such models. This requires extensive benchmarking against high-level electronic structure methods, including hybrid DFT, as well as ab initio molecular dynamics to capture finite-temperature effects and interfacial dynamics. These tasks are computationally demanding and drive the need for large-scale GPU resources. The first subproject focuses on redox-active metal oxides in aqueous environments, where the interplay between surface structure, defects, and interfacial water governs reactivity and functionality. The second subproject investigates molecular adsorption and interfacial chemistry in complex systems such as cellulose and graphene-based materials, with applications in sensing and environmental processes. Multiscale simulations are used to describe adsorption in the presence of water and heterogeneous environments. The third subproject addresses ion diffusion and structural ordering in solid-state electrolytes and battery materials. By constructing atomistic free-energy landscapes, the work links microscopic transport mechanisms to macroscopic performance. Together, these efforts combine method development with application-driven research to enable predictive modelling of complex materials systems.