Neuro-evolution is a sub-field of machine learning that takes an evolutionary approach to building neural networks. It has become more active again with the rise in popularity of deep neural networks. Research in the past year has shown that neuro-evolution methods can be applied in a number of novel manners, and can reach state-of-the-art accuracy of a number of classical tasks. While these are great results, they generally come at great cost in terms of computation time. This research focuses on improving the performance (both accuracy and computation time) of a particular neuro-evolution framework that can both build a network from scratch, and take an existing network and improve it. The research takes place at Philips Research under supervision of Dr. Sergio Consoli.