With the increased capability and speed of computing using GPU’s and HPC’s, industrial systems can now be studied in greater accuracy and depth than ever before. One process which has really benefited from these advances is vertical stirred milling; a complex system that requires large-scale DEM simulations, coupled with CFD, to predict fine particle grinding. Almost every industry relies on particle size reduction methods, with grinding from mining alone contributing almost 2% of global energy [1]. Therefore, a better understanding of these machines could improve efficiency and reduce energy consumption. The talk demonstrates how we can develop a validated mill by using GPU and HPC computing paired with experimental technique. Laboratory data collected using the Positron Emission Particle Tracking (PEPT) method shows that, despite being able to collect significant data, there are limitations with current physical techniques. This is where the model gives us an ability to assess the full design space. Once built, evolutionary algorithms were used to refine the material parameters until an optimal set of conditions were found. In certain cases, it was also found that the complex CFD-DEM coupling could be reduced or even removed due to the nature of the system.