Many natural and industrial processes involve particle-laden or droplet-laden flows where the dynamical evolution and transport of clouds of particles are intimately coupled with an underlying fluid flow in a carrier gas. Examples of such phenomena include volcanic eruptions, sand storms, pharmaceutical sprays, liquid fuel sprays, and particle-based solar receivers. In each application, the detailed physics can be tracked by simulating a large system, often requiring massive scale high performance computing resources. This is a challenging task from a computational science and computer science point of view. However, I would like to focus on the challenge of how to turn these large-scale simulations into simple learning lessons that could then guide future designers and engineers working on these applications. Often we seek simple quantitative models that are consistent with physical intuition but do not require significant calculations. I will discuss examples of such problems, and will demonstrate how in research the progress in large-scale computations is intertwined with the progress in simple understanding.