Yoshihiro Fukuhara

Data & Machine Learning

Career Summary

I bridge the gap between academia and industry. After researching computer vision at Waseda University and publishing at top-tier conferences such as CVPR and ECCV, I served as a specially appointed fellow at AIST. I also promoted industry–academia and international collaboration as an organizer of cvpaper.challenge and a core member of LIMIT.Lab, an international joint laboratory. After working in the CTO Office at ExaWizards, I joined CADDi in 2026. I now lead the research team, driving the social implementation of AI technologies in manufacturing, with a particular focus on 3D vision and multimodal recognition.

Why I Joined

My core mission has long been to turn knowledge born in academia into value for real industry. I had been watching the intersection of manufacturing and AI for some time, and the emergence of 3D foundation models in March 2025 convinced me that the field had reached a turning point. Advances in 3D and multimodal technologies made it possible for AI to understand and leverage the complex drawings, CAD data, and text information hidden within manufacturing, creating a foundation for full-scale real-world implementation. CADDi is a rare place where massive, complex manufacturing problems intersect head-on with world-class research themes. I decided to join because I was convinced it was the ideal place to connect cutting-edge research directly to an industry where Japan has a real global advantage—not by merely experimenting with technology, but by moving back and forth between papers, products, and customer value.

Role at CADDi

At CADDi, I lead the research team, driving the formulation and execution of a roadmap for foundational AI technologies that will serve as the source of our long-term competitive advantage. Specifically, I am developing core AI technologies that create immense value on the manufacturing domain by leveraging the latest advancements in 3D geometric and spatial recognition, multimodal processing, and knowledge graphs, combining them with CADDi's vast accumulation of diverse industrial data. Furthermore, utilizing my strong network with AIST and broader research communities, I propel technological development by actively integrating cutting-edge academic insights into our workflows. My ultimate mission is to highly fuse the perspectives of both researchers and engineers, ensuring that we generate globally competitive AI technologies directly from CADDi.