PCGRL via GA – Researcher and Prototype Dev – NYU Research Roles: Researcher and Prototype Developer of Procedural Content Generation via Genetic Algorithms research paper. Responsibilities: Co-research and main prototype developer for the research project. Research based on a previous study using Procedural Content Generation with Reinforcement Learning. Intention here is to replace RL with Genetic Algorithms as RL is a policy-gradient based approach which proved heavy to run. Translated this to a gradient free approach based on neural networks and genetic algorithms and based the research paper around this. Built neural network through a training process involving the modification of GVGAI levels and getting best chromosomes through the genetic algorithm. Used final evaluation to see how well best chromosome can build a playable level from a randomly generated one. Misc: The main intention of this research is to prove that there are more efficient alternatives to procedural content generation outside of Reinforcement Learning. We tested our methodology on chromosomes running on 15, 20, and 50 different generations and were able to record these results in our paper. The entire process is directly correlated to the study and practice of neuroevolution in Artificial Intelligence practices.