cv My CV

Member of Technical Staff, Intern

Amazon Frontier AI & Robotics

• Humanoids

Research Intern

General Vision Lab, BIGAI

• Humanoids

Research Intern

CoRe Lab, Peking University

• Humanoids
• CV4Animals

Bachelor

Peking University

Computer Science

• Machine Learning
• Computer Vision
• Robotics
• Reinforcement Learning
• Applied Stochastical Process
• Differential Manifolds

Bachelor

Peking University

Economics

• Microeconomics
• Macroeconomics
• Game Theory

Undergraduate Research Award

Yuanpei College, Peking University

Dean's Scholarship

Institute of Artificial Intelligence, Peking University

Yuanpei Scholar

Yuanpei College, Peking University

Soong Ching-ling Scholarship

Soong Ching-ling Foundation

Peking University Freshman Scholarship

Peking University

Finalist in International Science & Engineering Fair

International Science & Engineering Fair

Second prize in Tencent XINIUNIAO High School Science Award Competition

Tencent

First prize in China region, in S.-T Yau High School Science Award

S.-T Yau High School Science Award Committee

Action-Aligned Latent Planning for Versatile Humanoid Loco-Manipulation

arXiv

FIBER unifies humanoid loco-manipulation through action-aligned latent policy projection and offline perceptive planning, with real-world box pickup, door opening, object grasping, and skill chaining.

OmniClone: Engineering a Robust, All-Rounder Whole-Body Humanoid Teleoperation System

arXiv

A whole-body humanoid teleoperation system with robust multi-skill control, the OmniBench diagnostic benchmark, and subject-agnostic retargeting for scalable humanoid control and learning.

LessMimic: Versatile Humanoid–Object Interaction with Unified Distance Field Representations

Advances in Neural Information Processing Systems (NeurIPS) 2026

A whole-body humanoid policy conditioned on distance-field geometric cues for reference-free interaction, long-horizon skill composition, and vision-only deployment without motion capture.

GapONet: Nonlinear Operator Learning for Bridging the Humanoid Sim-to-Real Gap

NeurIPS 2025 Workshop on Embodiment-Aware World Models (EWM)

A payload-conditioned nonlinear operator that maps simulation context functions to residual actions for hardware, keeping large sim-to-real gaps below 0.09% when tracking unseen motions.

COLA: Learning Human-Humanoid Coordination for Collaborative Object Carrying

International Conference on Robotics and Automation (ICRA) 2026

COLA is trained in a closed-loop environment with dynamic object interactions to predict object motion patterns and human intentions implicitly, enabling compliant collaboration to maintain load balance through coordinated trajectory planning.

CLONE: Holistic Closed-Loop Humanoid Whole-Body Teleoperation for Long-Horizon Tasks

Conference on Robot Learning (CoRL) 2025

CLONE employs an MoE-based policy with closed-loop error correction for holistic humanoid teleoperation, enabling capabilities previously unattainable with existing systems—such as whole-body coordination and long-horizon task execution.

AlphaChimp: Tracking and Behavior Recognition of Chimpanzees

International Journal of Computer Vision (IJCV)

We propose AlphaChimp, an end-to-end approach that simultaneously detects chimpanzee positions and estimates behavior categories from videos.