Summary
Yongzhi Xu is a Senior Machine Learning Engineer and applied researcher based in Sydney with 11 years of experience building computer vision and spatial AI systems for industry leaders like TikTok, Tencent, and ByteDance. He leads teams that bridge research and production, with deep expertise in 3D reconstruction, visual-inertial SLAM, AR/VR positioning, and real-time scan-to-BIM workflows stemming from a PhD at UNSW. His work ranges from MEMS IMU calibration and EKF-based sensor fusion to neural approaches for vanishing point, line mapping, and single-view 3D surface detection, and he has been recognized internally (ByteStyle Awards) for impactful contributions. Yongzhi’s background in aerospace engineering and hands-on algorithm development (C++, Caffe, ORB_SLAM2, ARUCO) lets him move concepts from prototype to robust product features. He combines academic rigor—scholarship-backed PhD research and a Dean’s Award recommendation—with pragmatic leadership of Tencent’s 3D reconstruction team and ongoing ML engineering at TikTok. Notably, he often tackles the hard integration problems that make perception systems reliable in the wild, not just research benchmarks.
11 years of coding experience
5 years of employment as a software developer
Master of Engineering - MEng Inertial aided visual SLAM, Master of Engineering - MEng Inertial aided visual SLAM at Beihang University
Bachelor's degree Aircraft Design and Engineering, Bachelor's degree Aircraft Design and Engineering at Nanjing University of Aeronautics and Astronautics
UNSW Sydney
English