Resin Transfer Molding, or RTM, is one of the most popular methods in today’s world for manufacturing composite materials. The capacity for creating flexible and complex geometries of high volume fraction at lower costs makes it a very appealing manufacturing process for both the automotive and aerospace industries. However, no process is without its own drawbacks. One of the major issues faced in producing composites through this technique is the inclusion of voids and air bubbles. Naturally, these act as points of failure or low strength in the final created part. The first contribution of this thesis deals with identifying void formation mechanisms through simulations of the RTM filling process in commercial software (ANSYS FLUENT). This is followed by the computational exploration of resins that exhibit Non-Newtonian behavior and how they affect void formation. Finally, we use Machine Learning Optimization tools to determine the best set of resin rheology parameters from the conceivable sample space, and study the corresponding effects on fill. Through this innovative approach, we are able to establish a simulation framework for RTM, a unique fluid-based approach to tackle an industry-wide problem and a deeper understanding of how certain material properties can be cultivated to our advantage.