Quiet mind, sharp work.

Hi, I am

Li Zhaoyang

3D Perception · Robotics · Embodied AI

I study how machines recover geometry, reason across modalities, and move through the world, from point cloud completion to embodied robotic systems.

Open to internships CV INCOMING
Publications

SplAttN: Bridging 2D and 3D with Gaussian Soft Splatting and Attention for Point Cloud Completion

Zhaoyang Li, Zhichao You, Tianrui Li

ICML 2026 Spotlight Top 2.2%

PaCo-VLA: Passivity-Shielded Compliance Prior for Contact-Rich Vision-Language-Action Manipulation

Haofan Cao, Zhaoyang Li, Zhichao You, Liang Guo, Tianrui Li

ICML logo

Robotics Showcase

My Current Work in Robotics

Interactive demonstrations from my ongoing robotics work, covering VLA/WAM-based robotic control and mobile manipulation system integration. These showcases reflect my practice in robot modeling, kinematic visualization, manipulation workflows, mobile-base assembly, and embodied AI-oriented system design.

Showcase 01 VLA/WAM-based Robotics Control A manipulation-oriented control showcase connecting visual perception, language-conditioned task logic, and robotic action execution.
Showcase 02 Mobile Manipulation System Integration A Scout-AUBO mobile manipulator platform for exploring arm-base coordination, embodied robot assembly, and future navigation-manipulation tasks.
Showcase 01

VLA/WAM-based Robotics Control

VLA · WAM · Manipulation Control
Showcase 02

Mobile Manipulation System Integration

Arachne repo AgileX Scout 2.0 · AUBO i5

Education

Academic Path

Southwest Jiaotong University

PhD Student, Computer Science

Advisor: Prof. Dr. Tianrui Li

Southwest Jiaotong University

M.Sc. Student, Computer Science

Advisor: Prof. Dr. Tianrui Li

Centrale Méditerranée

Generalist Engineering student

Southwest Jiaotong University

Undergraduate, Electronic Information Engineering

Language Ability

Multilingual Training

Credentials across English and French academic environments.

CET-4 English
CET-6 English
TOEIC 920 English
DELF B2 French

Coding Ability

AI-Native Development

Practical engineering now starts with agentic coding systems and fast iteration.

GitHub language profile

Aggregated from public zay002 repositories

Primary workflow AI-native research engineering

Fast prototype loops with Codex, Claude Code, Python, CUDA, and web demos.

Codex Claude Code
Most used languages Auto-updated daily
Python60.1%
JavaScript18.3%
TypeScript14.4%
HTML2.2%
CUDA1.9%
CSS1.7%

Contact

Email

Open to research collaboration, internships, and full-time opportunities.