Our concordance cosmological model, known as Lambda-CDM, is remarkably successful in explaining a wide range of observations. Using data to fit that model to we are able to constrain its parameters. In the era of precision cosmology, we have reached a point where we can constrain those parameters remarkably well and as we move into the next generation of galaxy surveys, we expect to be able to do even better. However, this also means that we need to build models that keep up with the increasing precision of the data. Moreover, we need to be able to utilize the whole data set, which we are not capable of doing well enough presently. Specifically, cosmological studies from galaxy surveys by collaborations such as the Dark Energy Survey (DES) only use the large, linear scales due to our inability to accurately model the non-linear regime. The removed scales, however, contain a wealth of information that could lead to increased constraining power on cosmology, as well as on the physics of how galaxies relate to their dark matter halos, which is a key component of the Lambda-CDM model. Furthermore, the Lambda-CDM model itself can be stress-tested on the non-linear scales, which can reveal more about the nature of the dark matter and even put General Relativity to test. It becomes clear, therefore, that we need to model small scales if we want to construct a complete and comprehensive picture of the cosmological model of the Universe. In this thesis, we focus on developing galaxy-halo connection models in order to model the non-linear scales in 2-point statistics in Cosmology. We use our framework to analyze data from the Year 3 (Y3) of the DES, self-consistently describing the large scales used in cosmological studies and the small scales that probe the connection between galaxies and the dark matter distribution in the Universe. We successfully obtain constraints on the relation between galaxies and their dark matter host halos in DES Y3 at fixed cosmology. Our model can be utilized in the future to perform studies where, in contrast, we fix this "galaxy-halo connection" to model all scales in the 2-point statistics of interest to constrain the cosmological parameters, or where both sets of parameters vary at the same time.