curriculum vitae
Basics
| Full Name | Heqing Yang |
| Affiliation | School of Artificial Intelligence, Beihang University |
| Advisor | Prof. Si Liu (co-advisor Prof. Linjiang Huang) |
| yangheqing@buaa.edu.cn | |
| Location | Beijing, China |
Education
- 2024 – Present
Ph.D. in Artificial Intelligence
Beihang University, Beijing
- Advised by Prof. Si Liu; co-advised by Prof. Linjiang Huang.
- 2019 – 2024
B.S. in Software Engineering
Beihang University, Beijing
- GPA 3.93/4.0 (rank 8/187).
Experience
- 2025.10 – present
Research Intern
Meituan LongCat Team
- Robot foundation models.
- 2024.03 – 2025.01
Research Intern
BIGAI (Beijing Institute for General Artificial Intelligence)
- Long-horizon robot task planning on 3D scene graphs — first-author ICRA 2026 submission (EPoG).
- 2020.12 – 2023.09
Team Leader — BRBOTS humanoid biped project
BUAA Robotics Team
- 18-DoF humanoid biped built end-to-end across two hardware iterations (mechanical, electronics, firmware, software).
- 2nd Prize (top-4), RoboCup Asia-Pacific Championship.
Selected Projects
-
EPoG — Long-horizon task planner on 3D scene graphs
- Graph-edit planner with LLM-guided heuristic that fuses task- and motion-level reasoning; 91.3% success rate on 5 long-horizon transport tasks across 46 ProcThor-10k scenes.
- Validated on a physical mobile manipulator handling unknown object states and runtime exceptions in real, dynamic environments.
-
CR-Solver — GPU-accelerated kinematics for continuum robots
- Unified nonlinear-optimization framework for IK, path following, and trajectory planning; pure Python/JAX with GPU parallelism.
- > 95% success rate with millimeter-level accuracy.
-
EVA-Client — Unified framework for deployment, evaluation, and data collection on real robots
- One policy, any robot — a smooth all-in-one real-robot stack. Debug, record, evaluate, and visualize entirely in the browser.
- Drives an AgileX bimanual arm end-to-end (teleop → record → π₀ checkpoint → smooth async deploy) on real hardware.
-
BRBOTS — 18-DoF humanoid biped robot
- Team leader across two hardware iterations; hardware, electronics, firmware, ROS stack, gait planner, particle-filter visual localization.
- 2nd Prize (top-4), RoboCup Asia-Pacific Championship.
Awards
- 2nd place (2/62), RoCo Challenge @ AAAI 2026, 2026
- 1st Prize, RoboCon, 2021 & 2022
- 2nd Prize (top-4), RoboCup Asia-Pacific Championship, 2021
- Finalist, U.S. MCM/ICM Mathematical Contest in Modeling, 2021
Research Interests
- Embodied AI · Robot task and motion planning
- Open-vocabulary perception
- Continuum / soft-robot kinematics
- Active perception for legged robots