The typical method of design and development of process industry involves thorough experimentation at three levels: lab scale, pilot scale level and industrial level. This scale requires huge investment for process evolution. In past few decades, computational fluid dynamics (CFD) has emerged to be a relatively lowcost alternative. Being a physics-based approach, CFD is a reliable tool for scientific research. However, keeping in view the computational requirements, CFD may take a lot of computational time and resources for model development and validation. In this respect, the model development may require accurate inputs in terms of physical properties and boundary conditions. This is where machine learning may be coupled with CFD to identify parameters for global process development. Once a regression model is fitted, an optimization algorithm may be used to identify the best and worst case scenario making the operating and geometric parameters available. These parameters can then be used to perform comprehensive CFD modeling of the process in question. The results thus obtained will be helpful at all scales and assist designers to reduce the number of experiments hugely. This understanding can reduce the time of process development, new material discovery and save precious financial resources.