This research project was undertaken at the Summer Lab Program of the UChicago Data Science Institute. Its goal was to build a pipeline for researching Federated Machine Learning models for computer vision. Federated learning involves building a machine learning model that learns from data scattered across multiple devices, with the constraint that this data cannot be shared between them. This requires aggregating model parameters at different levels to combine the characteristics learned in each of the local devices eventually combining them into a single global model. This global model thus benefits from all the scattered data while preserving its privacy and security. More specifically, I focused on researching how the process of aggregation affects the resulting model structure and performance.