This proposal focuses on the integration of two different models for semiconductor heterostructures with different characteristic length scales. The electronic bandstructure of semiconductor superlattices is commonly obtained from self-consistent Schrödinger-Poisson models, which are based on the envelope function approximation and therefore use a continuum-based approach for the relevant material parameters. Both, microscopic effects like interface roughness scattering and long-range effects like graded structures, or other growth-induced imperfections are commonly neglected in the electronic model. Within this project, the k.p formalism is used to obtain bound states and band-dispersions of heterostructures to provide accurate scattering rates. Due to the length scales involved, this model can not directly account for interface roughness scattering, which previously was considered in form of fixed rates. Therefore, an atomistic model is used to simulate the formation of interface roughness during epitaxial growth as a complementary technique. Empirical interaction potentials simplify the description of chemical bonds and therefore allow large simulation boxes – beyond the capabilities of ab-initio models – which still keep the crystalline structure of semiconductor materials. A Metropolis Monte Carlo-based algorithm is then used to optimize the bond configuration at the interfaces. Resulting roughness features shall be integrated with the electronic model, composed of a k.p-based model to determine the dispersion and scattering rates, which are then input into a rate-equation model to describe domain formation, fluctuations and long-range effects within the active region superlattice. This combination will allow a more realistic description of interface roughness scattering in heterostructure-based devices like resonant tunneling diodes or quantum cascade lasers.