The transition towards sustainable energy technologies requires new functional materials with tailored properties. Materials modelling has become an indispensable tool in this process by providing atomic-scale insight into materials behaviour and guiding experimental efforts. The principal investigator has extensive experience in ab initio modelling of functional materials [1-3]. This project addresses two complementary research directions that require large-scale high-performance computing resources. 1) Electronic structure calculations and x-ray absorption spectroscopy (XAS) of Ce-based materials. The work is part of the MSCA EUSpecLab . Cerium compounds are particularly challenging because of the mixed-valent character of the 4f electrons, making accurate theoretical descriptions demanding. We will perform systematic calculations of XAS and, where appropriate, x-ray magnetic circular dichroism (XMCD) using the DFT+DMFT implementation in the RSPt code together with multiple-ligand field theory [4,5]. In parallel, we will benchmark the multiple-scattering code SPRKKR [6] for Ce-based compounds against RSPt calculations and experimental data. SPRKKR is advantageous for disordered systems because supercell calculations could be avoided. In collaboration with the University of Duisburg-Essen, we will investigate Ce2Fe14B-based compounds to understand how rare-earth substitution influences the localization of the Ce 4f electrons and how these changes are reflected in the XAS and XMCD spectra. 2) Computational design of next-generation rare-earth-lean permanent magnets based on REFe12. These materials exhibit excellent intrinsic magnetic properties but require chemical substitution to stabilize the crystal structure. We will perform high-throughput ab initio calculations of magnetization, magnetocrystalline anisotropy, and exchange interactions using RSPt, followed by atomistic spin-dynamics simulations with UppASD [7] to determine finite-temperature magnetic properties. To efficiently explore the large compositional space, machine-learning interatomic potentials will be employed to identify favourable dopant configurations prior to DFT calculations. Their predictive accuracy will be validated using VASP calculations [8]. This work contributes directly to the European project MaMMoS, which aims to accelerate the discovery of functional magnetic materials through multiscale computational modelling. Both parts require extensive electronic-structure calculations plus finite-temperature simulations, and systematic exploration of chemically complex materials. Access to Arrhenius through NAISS is therefore essential to achieve the scientific objectives within a reasonable timeframe. [1] H. C. Herper, Physical Review B 98, 014411 (2018) [2] H. C Herper et al., Acta Materialia 242,118473 (2023) [3] H. C. Herper et al., Physical Review Materials 1, 033802 (2017) [4] J. M. Wills et al., Full-Potential Electronic Structure Method, Vol. 167 of Springer series in solid state science (Springer, Berlin, Germany, 2010). [5] F. Sorgenfrei et al., Phys. Rev. B 109, 115126 (2024). [6] The Munich SPR-KKR package, version 9.7, H. Ebert et al., https://www.ebert.cup.uni- muenchen.de/sprkkr. [7] O. Eriksson et al., Atomistic Spin Dynamics: Foundations and Applications (Oxford University Press, Oxford, 2017). [8] G. Kresse and J. Hafner, Phys. Rev. B 49, 14251 (1994), G. Kresse and J. Furthmüller, Comp. Mater. Sci. 6, 15 (1996).