Fangqi Zhu

Email Google Scholar

HKUST CSE PhD · NVIDIA GEAR intern · 2027 industry job market

I am a second-year PhD student in the Department of Computer Science and Engineering at HKUST, advised by Prof. Song Guo at PEILab. I am broadly interested in robot learning, world models, and reinforcement learning.

My current research explores how robots can learn from models of the physical world, with the goal of making robot learning more scalable, data-efficient, and adaptable across tasks and environments.

Since May 2026, I have been a research intern at NVIDIA GEAR, working on world models for robot learning with Shenyuan Gao, Yuke Zhu, and Jim Fan. I expect to graduate in 2027 and am on the industry job market.

2026.05

Started at NVIDIA GEAR

Research internship with Shenyuan Gao, Yuke Zhu, and Jim Fan on world models for robot learning.

2026.05

HALO accepted to ICML 2026

Work on embodied multimodal chain-of-thought reasoning and action generation.

2026.01

WMPO accepted to ICLR 2026

World model-based policy optimization for robot learning.

2025.06

IRASim accepted to ICCV 2025

Interactive real-robot action simulators for generative robot world models.

HALO method overview

ICML 2026

HALO: A Unified Vision-Language-Action Model for Embodied Multimodal Chain-of-Thought Reasoning

Quanxin Shou*, Fangqi Zhu*, Shawn Chen, Puxin Yan, Zhengyang Yan, Yikun Miao, Xiaoyi Pang, Zicong Hong, Ruikai Shi, Hao Huang, Jie Zhang, Song Guo

* Equal contribution.

HALO combines textual reasoning, visual subgoal prediction, and action generation for embodied multimodal chain-of-thought reasoning.

Paper Code

WMPO overview

ICLR 2026

WMPO: World Model-based Policy Optimization for Vision-Language-Action Models

Fangqi Zhu, Zhengyang Yan, Zicong Hong, Quanxin Shou, Xiao Ma, Song Guo

Explores model-based policy optimization in a learned world model.

Paper Code Project

IRASim overview

ICCV 2025

IRASim: Learning Interactive Real-Robot Action Simulators

Fangqi Zhu, Hongtao Wu, Song Guo, Yuxiao Liu, Chilam Cheang, Tao Kong

Studies generative robot world models for interactive real-robot action simulation.

Paper Code Project

Earlier NLP publications EMNLP 2023, ACL Findings 2023, AAAI 2023

DDI paper overview

EMNLP 2023 · Oral

Learning to Describe for Predicting Zero-shot Drug-Drug Interactions

Fangqi Zhu, Yongqi Zhang, Lei Chen, Bing Qin, Ruifeng Xu

Paper Code

ACL Findings 2023

Event skeleton generation overview

ACL Findings 2023

A Diffusion Model for Event Skeleton Generation

Fangqi Zhu, Lin Zhang, Jun Gao, Bing Qin, Ruifeng Xu, Haiqin Yang

Paper Code Poster

Script event prediction overview

AAAI 2023 · Oral

A Generative Approach for Script Event Prediction via Contrastive Fine-Tuning

Fangqi Zhu, Jun Gao, Changlong Yu, Wei Wang, Chen Xu, Xin Mu, Min Yang, Ruifeng Xu

Paper Code Poster

2026.05 - Present

NVIDIA GEAR

Research Intern · World models for robot learning

Collaborating with Shenyuan Gao, Yuke Zhu, and Jim Fan.

2025.09 - 2026.05

ByteDance Seed Multimodal

Research Intern · Multimodal foundation models

Mentor: Haoqi Fan.

2024.03 - 2025.09

ByteDance Seed Robotics

Research Intern · Robot world models and real-robot simulation

Mentors: Tao Kong, Hongtao Wu, and Xiao Ma.

2023.04 - 2023.12

4Paradigm

Research Intern · NLP and drug-drug interaction prediction

Mentor: Yongqi Zhang.

2022.09 - 2023.03

International Digital Economy Academy

Research Intern · DataStory AI Lab

Mentor: Haiqin Yang.

2024 - Present

PhD in Computer Science and Engineering

Hong Kong University of Science and Technology · Advisor: Prof. Song Guo

2021.09 - 2024.01

Master of Engineering in Computer Science and Technology

Harbin Institute of Technology, Shenzhen

2017.09 - 2021.06

Bachelor of Engineering in Software Engineering

Wuhan University