In this talk I will describe two complementary and novel approaches to understanding the assembly of galaxies beyond the MW in a cosmological context using observational tracers. In the first approach, we infer the origin of globular clusters from observables towards the goal of reconstructing the assembly histories of galaxies in upcoming wide-field surveys. Using the E-MOSAICS simulations to follow the formation and co-evolution of ca. 1000 galaxies and their star clusters, we explored the use of supervised machine learning to classify observed GCs into accreted and in-situ populations. Assessing the performance using a subset of the simulations and the known origin of the MW clusters, we obtain an accuracy of ca. 90% for 2/3 of the sample and successfully identify accreted debris buried deep within the Galaxy. In the second approach we study hundreds of high quality galaxy rotation curves to understand the impact of the large-scale environment on the structure of their host dark matter (DM) haloes. Galaxies in high density environments show a systematic shift in their DM density profile at large radii that is consistent with a relatively early assembly of their host haloes. The effect is manifest in the well known radial acceleration relation (RAR) as a slight downturn at the lowest accelerations for galaxies in dense environments. This environmental dependence can be understood in the context of assembly bias within the Lambda-CDM cosmological paradigm, implying that the RAR can provide useful constraints on the fundamental relation between dark and luminous matter.