Advances in computational methods in recent decades have significantly expanded the range of problems in condensed matter physics that can be tackled from first principles. Linear-scaling density functional theory methods enable quantum mechanical calculations to be performed on systems containing tens of thousands of atoms, with modern approaches capable of reproducing the accuracy of plane wave DFT approaches. This opens up the possibility of treating highly complex molecular systems such as doped organic molecular crystals that require the dopant molecule to be contained within a large periodic structure. One example of such a system is pentacene in p-terphenyl, a system that finds use as a room-temperature maser. Understanding the maser mechanism requires both a highly accurate description of the pentacene molecule and a computationally efficient approach that can correctly capture the impact of the p-terphenyl host on the active pentacene subsystem. Quantum embedding allows an accurate but expensive hybrid functional to be embedded within a cheaper semi-local functional, for maximum combination of accuracy and efficiency in a DFT-in-DFT framework. In this dissertation we consider the implementation of embedded mean-field theory (EMFT) in the linear-scaling DFT software package ONETEP, enabling hybrid functionals to be used on selected subsystems within a cheaper DFT environment. This approach is validated for several types of molecular systems, including a crystalline structure containing several thousand atoms, demonstrating the potential of the EMFT approach when combined with linear-scaling and verifying the importance of using a large explicit host environment for accurate calculations.