The cosmic origin of elements heavier than iron remains one of the open questions in astrophysics. These elements are produced through neutron-capture processes in diverse astrophysical sites, but, to disentangle their contributions and galactic evolution, we rely on accurate stellar abundances. Deriving these abundances requires reliable non-LTE models of stellar spectra. However, for many heavy elements, non-LTE modelling is still missing due to the scarcity of accurate atomic data. This challenge highlights the interconnected nature of atomic physics and astrophysical interpretation: without reliable atomic data, we cannot fully exploit the wealth of stellar observations.

I have worked at this intersection, aiming to improve atomic data and construct more accurate non-LTE models for heavy elements. I will present results from different stages of this work: from theoretical calculations of atomic structure and transition probabilities to non-LTE modelling of copper and silver. I will show how improved atomic data, such as hydrogen collision rates and photoionisation cross-sections, impact non-LTE abundance determinations of these elements and how refined abundance trends in turn sharpen our understanding of nucleosynthetic origins and galactic chemical evolution.