Member
Development Team
University of Tokyo Engineering / Geotechnics x Physics-Informed AI
University of Tokyo Engineering / Geotechnics x Physics-Informed AI
Hiroki has led web development businesses from sales and requirements definition through implementation, delivery, and post-delivery improvement. He also experienced founding and technical leadership at an AI development and consulting company backed by deep30, a venture capital firm formed by members of the University of Tokyo Matsuo Lab community.
As a software engineer, he has worked on Kotlin and Flutter app development while conducting research on deep learning models including speech enhancement. His work spans loss-function design, fine-tuning comparisons, signal-processing pipelines, and real-world audio evaluation.
As CEO, he leads customer communication, business development, and product strategy for practical physics-informed AI.
After receiving a Ph.D. in geotechnical engineering from Saitama University, Zafar joined the University of Tokyo Graduate School of Engineering as a project researcher. His work started with the physical mechanisms of rainfall-induced landslides and debris flows, and expanded into physics-informed machine learning and glacier retreat prediction.
His paper on the three-dimensional dynamic behaviour of embankments on liquefiable ground was published in Geotechnique Letters and received the ICE Telford Premium Prize. He has also released multiple GitHub repositories implementing physics-informed machine learning and geotechnical analysis models.
As CTO, he leads the integration of physical models and AI for predictive maintenance under data scarcity and slope collapse risk quantification.