Nomi Yu

Nomi Yu is a graduate student in mechanical engineering, developing generative AI tools to accelerate design and reverse engineering in the manufacturing industry. Their research focuses on reconstructing and generating editable CAD programs from non-parametric geometries, like point clouds and meshes, enabling engineers to remanufacture legacy parts and refine AI-generated geometries. As a second-year MathWorks Fellow, Nomi will extend this work into a streamlined, end-to-end pipeline for scan-to-CAD reconstruction and automated design analysis. Their framework, GenCAD-3D, combines modality-specific encoders, contrastive learning, and diffusion-based generation to produce high-fidelity CAD feature trees from point clouds and meshes. They will use MATLAB for integrating scan preprocessing, model training, and downstream simulation, leveraging toolboxes for lidar, computer vision, deep learning, and CAD interfacing. Nomi’s work could improve accessibility to generative design tools, reduce engineering cycle times, and expand the practical use of AI in product development across robotics and advanced manufacturing.