INDUSTRY AFFILIATIONS
Zuru generates multi-sensor synthetic training data — EO, depth, and thermal — from 3D digital twins, enabling AI perception systems to train, validate, and deploy in air-gapped, classified, and data-denied environments without a single real-world sensor collection.
Zuru's first paid commercial deployment — automotive remanufacturing — demonstrates the full capability stack at TRL 7: multi-sensor synthetic data generation, model training, and real-world validation. No real-world sensor data collected.
Matthew holds a BA in Sociology from Stanford University and an Executive MBA from INSEAD's Global Executive MBA program.
He led North American business development at two computer vision AI startups — working at the front lines of how perception AI gets sold, deployed, and where it consistently stalls. The bottleneck was never the model. It was always the training data.
Yu-Feng holds a PhD in Mechanical Engineering from MIT and serves as the architect of Zuru's synthetic data generation method — the core ML technology underlying the platform.
Over two decades, he has deployed machine vision AI across manufacturing, clinical imaging, and industrial inspection systems. He served in the Republic of China Marine Corps, bringing direct military operational context to the defense environments Zuru serves. That breadth — simulation, sensor physics, real-world deployment — is uncommon at the technical founding level.
Zuru's core software development is conducted on U.S. soil by our founding engineering team.
Real-world training data is scarce and impossible to collect at scale — especially in defense, energy infrastructure, and industrial environments where sensor data is classified, operationally restricted, or simply unreachable in air-gapped and denied-access environments.
We're building the data infrastructure that makes autonomous systems deployable at scale — in defense, energy, and industrial environments where real-world data cannot be collected.
