Atomistic simulations provide a necessary lens through which to characterize nanoscale phenomena. This dissertation begins with a description of molecular models and the development of aninteratomic potential for benzene which incorporates atomic-level anisotropy. This model was made possible for bulk benzene systems through the implementation of a software plugin for the OpenMM simulation package, which enables custom force expressions with atomic-level anisotropy. This initial discourse on force field development summarizes the types of interatomic potentials used in simulations and avenues for improved accuracy. This knowledge of fundamental force field development is transferrable to developing approaches in modeling inorganic crystallization. Solid-phase epitaxy (SPE) is a crystal growth technique which employs low-temperature annealing conditions to exact kinetic control over the final grown structure. In this dissertation, classical simulations are used to rigorously define the mechanism of epitaxial growth in strontium titanate over patterned substrates. Modeling SPE is challenging from a simulation perspective because long timescales at experimental growth temperature exceed computational feasibility. The enhanced sampling method, metadynamics, is presented here as a viable alternative for probing crystallization mechanisms in super-cooled and viscous systems, for which diffusion is limited. Gaining mechanistic information from metadynamics is dependent on the "goodness" of reaction coordinate. Here, an XRD-based coordinate is used to distinguish not only between the amorphous and crystal structures but also among metastable crystal polymorphs. This dissertation summarizes work which encompasses research spanning molecular models and inorganic crystallization with added commentary on outreach and communication.