Colloidal nanoparticles at the interface of two immiscible fluids experience certain restrictions on their position and orientation. This property is exploited to create long-range, coherent assemblies of quasi-2D super-structures that are known to possess strong correlations between their packing symmetries (structure) and the displayed opto-electronic properties (function). However, despite the recent advances in synthesis techniques, the underlying kinetic and thermodynamic factors governing the self-assembly process are not yet completely understood. The overarching goal of this work is to increase the repeatability, precision, and control over the self-assembly of constituent NPs into superstructures with programmable symmetry. In this work, I will take you through not only 1) the development of a set of design rules based on energetic arguments obtained from simulations and theoretical considerations, but equally importantly 2) the development of a simulation paradigm that is faithfully able to reproduce the inherent physics, in-silico. The first step in this process is to investigate the behavior of an isolated NP at the interface. For this purpose, I use both particle-based coarse-grained molecular simulation and a theoretical continuum model. I present the free-energy characteristics of the NPs as a function of their orientations and their vertical positions with respect to the interface. Meaningfully probing the self-assembly process at meso-scales requires simulation of O(10^3) NPs. However, this is infeasible in an explicit-solvent setting with the typically available computing resources. To this end, a key contribution of this work is to develop an efficient (implicit-solvent) model that is not only able to reproduce experimentally exhibited behavior by NPs at fluid-fluid interfaces but is also scalable to the experimentally relevant length scales. By explicitly modeling coarse-grained ligands that cap the nanoparticle surface, I show that changes in nanoparticle shape and ligand densities give rise to drastically different mechanisms. In agreement with experiments, formation of bilayer honeycomb and monolayer square lattices is observed. My results indicate that the choice of solvent and rate of evaporation have a significant impact on reversibility and ultimately the coherence of the finally obtained superstructure. The proposed simulation paradigm would pave the way forward for exploration of the vast phase space.