About Me


Zhengyan Lambo | BEng Student in Computer Engineering -- HKUST

Research Focus:​ 
Advancing general-purpose robotics through unified sensor representations, data-efficient learning, and tactile intelligence. My work lies at the intersection of robot learning, representation learning, and human-computer interaction, with the goal of creating robots that learn and adapt through experience, much like humans.

Biography
Zhenyan's journey into robotics began at age 14, building a boxing robot from plywood and servos. The stark gap between simple actuation and true, understanding-driven action sparked a lasting fascination with embodied intelligence. This early failure framed a core insight: true mastery, whether in badminton or robotics, comes not from pre-programming but from iterative, experiential learning.

He is now interested to investigate how to standardize robotic state representations across different hardware platforms. A key challenge is the lack of standardization in tactile sensing; each sensor type produces unique data, hindering the development of large-scale, trainable models. 

Zhengyan developed a diverse research portfolio at HKUST and ASTRI, demonstrating a consistent ability to apply interdisciplinary thinking to complex problems:
  • Medical Imaging (ASTRI):​ He conceived a novel, physics-inspired diffusion algorithm to solve a critical "double surface artifact" in 3D stomach model reconstruction, leading to a submitted first-author paper at CVPR.
  • Robotic Manipulation:​ He built a complete tactile feedback system for the delicate task of cable insertion, developing data collection frameworks and implementing an LSTM-based control strategy.
  • Human-Computer Interaction:​ He contributed to "Swift," a system for on-hand handwriting recognition for smartwatches using acoustic sensing, developing the data pipeline and achieving high recognition accuracy on unmodified hardware.
Research Interests
  • Embodied Artificial Intelligence
  • Robot Learning (Reinforcement Learning, Self-Supervised Learning)
  • Tactile Sensing and Perception
  • Universal Sensor Representation Learning
  • Human-Robot Interaction
Education
  • BEng / MSc in [Computer Engineering]​ Hong Kong University of Science and Technology (HKUST)
Selected Publications
  • Zhengyan Lambo Qin., Qiu, L., et al. "Diffusion-Driven Inter-Outer Surface Separation for Non-Watertight Point Clouds." Submitted to CVPR 2026.
  • Wentao Xie., Zhengyan Lambo Qin., et al. "SkinWriter: Transforming the Hand into a Text Input Method on Unmodified SmartWatches." Submitted to CHI 2026.

YouTube Channel: link

Bilibili ChannelL: link

GitHub: link