Haozhe Zhu · Researcher · Max Robotics

About

Haozhe Zhu is a rising star at the intersection of computer architecture, 3D vision, and robotics, pioneering hardware-software co-design for next-generation Simultaneous Localization and Mapping (SLAM) systems. His research focuses on creating specialized processors that enable real-time, energy-efficient SLAM on edge devices—a critical challenge for autonomous robots, drones, and augmented reality. Zhu’s major contributions include GauSPU, the first dedicated processor for 3D Gaussian Splatting (3DGS) in SLAM, achieving high-fidelity reconstruction and rapid rendering (17 citations). He also introduced SLAM-CIM, a backend processor featuring dynamic-range-driven-skipping and compute-in-memory macros for ultra-low-latency linear solving (10 citations). His latest work, RT-FLOW, demonstrates an FPGA-based optical-flow SLAM system for high-speed tracking and mapping. With a total of 29 citations across his most-cited papers, Zhu is establishing himself as a leader in hardware acceleration for visual intelligence. His innovations directly address the resource constraints of real-world deployment, making him a key figure to watch in the evolution of autonomous systems.