Designing robust manufacturing processes and avoiding trial and error methods are key for economical implementation of fiber reinforced plastics. Usually, this is achieved using process simulation based on the finite element or finite volume method. However, when simulating the fluid flow through a fibrous structure the inherent complexity, more precisely the interaction of two different materials (fiber and polymer) as well as the influence of physical effects e.g. intermolecular forces that act at different spatial and temporal scales, lead to major challenges. Here, Machine Learning could help to provide computationally efficient simulations by e.g. data driven surrogate modelling, coupling ML with conventional flow solvers, Model Order Reduction or Physics Informed Neural Networks. The goal of the project "ML4ProcessSimulation - Machine Learning for Simulation Intelligence in Composite Process Design" is to explore the aforementioned methods in order to develop a multiscale simulation workflow that reduces the computational effort and accounts for relevant physical effects.