The rise of AI in engineering design and manufacturing has accelerated development cycles and reduced costs, but its full potential is limited by fragmented and sensitive datasets spread across organizations. Federated learning offers a solution by enabling collaborative model training without exposing proprietary data. This project focuses on developing and implementing federated learning models suitable for engineering applications. I will evaluate different federated learning architectures and integrate techniques such as horizontal federated learning, secure model aggregation, and synthetic data generation to enhance efficiency and data security. The outcome will be open source methods and practical guidelines to support secure, collaborative manufacturing workflows.

I am participating in SuperUROP because I’m interested in how AI can assist in the design process. I’ve previously worked on a project where I generated a synthetic dataset of airfoils intended to train machine learning models for optimizing aerodynamic design. I’m excited to build on this foundation and explore how machine learning can drive innovation in engineering design and manufacturing.