Chenhao Zhang
About Me
Hi there! I am a senior undergraduate at Huazhong University of Science and Technology (B.S. in Intelligent Science and Technology, 2023 - 2027), originally from Taipei, Taiwan. I am currently a research intern at Tencent Hunyuan with Tianyu Pang, and at VisionX Lab, NYU Courant with Prof. Saining Xie, where I spent Fall 2025 as an exchange student. Previously I worked at OpenDILab, Shanghai AI Lab with Yazhe Niu, and at CAS-SIAT with Prof. Min Yang.
I am applying for CS PhD programs starting Fall 2027 and exploring full-time research opportunities at frontier AI labs worldwide. My research focuses on vision-centric multimodal perception and reasoning, Agentic Vision & Multimodal Coding, and multimodal RL and self-improvement. I also welcome long-term research internships and collaborations in these areas.
💬Research Interests
- Vision-Centric Multimodal Perception & Reasoning: Long-horizon visual reasoning, video and spatial perception, visual representation learning, visual state tracking and memory, and predictive world models.
- Agentic Vision & Multimodal Coding: Thinking with images, tool-augmented pixel operations, explicit visual states, Visual2Code, Code2Video, inverse graphics, visual harnesses, and visual verification.
- Multimodal RL & Self-Improvement: RL post-training, policy optimization, reward and verifier design, and verifier-guided self-improvement of agent policies and harnesses.
- Emotional Intelligence & Human-Centered AI: Mental world models, emotion reasoning for decision-making, emotion as long-horizon state compression, and evaluation across social, cultural, and mental-health domains.
I work on the gap between what a model sees and what it understands: implication, spatial relationships, and how states change over time. I believe strong visual representations are fundamental to reasoning, generation, and action.
My focus is Agentic Vision & Multimodal Coding, with coding agents as the primary mechanism. My bet is that executable visual environments can provide vision with what code offers language models: a training ground where agents work with real scenes over long horizons, verify outcomes, and learn transferable skills. A vision agent gathers evidence, uses visual tools, and maintains and revises visual state. Through coding and inverse graphics, it turns that understanding into executable scene programs, then renders, inspects, and refines the results. This unifies visual understanding and generation within a single agent.
I believe inspectable states and verifiable feedback offer a path to scaling both learning and inference. The challenge is to connect learned representations, predictive world models, and agent systems so that improvements generalize across tasks and environments.
Long-horizon intelligence must also account for human goals. I am interested in emotion as an experience-shaped value signal that guides attention, memory, and decisions. My goal is to build vision-centric agents that see, remember, reason, create, and act reliably in the real world.
🔥News
- Aug 2026 VisHarness, the project I lead, has been released as a lightweight visual agent harness for VLMs, with 25/25 games · 183/183 levels · score 100.0 on ARC-AGI-3 and 7,340 official actions — far fewer than humans' 17,135 actions!
- Aug 2026 HyCreator, the project I core contribute to, has already been released; check it out!
- Jun 2026 I will attend ICML 2026 @Seoul, please feel free to connect with me!
- Apr 2026 I have joined Tencent Hunyuan in Shenzhen as a Qingyun Talent Program Research Intern, working on multimodal RL and long-horizon visual agent!
- Apr 2026 OpenDILab, the project I core contribute to, already has 30K+ Stars!
- Mar 2026 Our paper Let Androids Dream of Electric Sheep has been accepted to CVPR 2026 CogVL, check it out!
- Feb 2026 I'm honored to be invited to be the Qinyuan member of Beijing Academy of Artificial Intelligence (BAAI)!
- Feb 2026 We preprint MetaphorStar on arXiv, check it out!
- Oct 2025 I have arrived in NYC and started an on-site research internship at NYU Courant, advised by Prof. Saining Xie. Looking forward to the exciting research ahead!
- Oct 2025 I'm so happy to win Baosteel Scholarship 2025!
- Jun 2025 I'm so happy to be 1/30 of SenseTime Scholarship 2025! Thank you, SenseTime!!!
- May 2025 I'm honored to be invited to be a reviewer of EMNLP 2025!
- May 2025 We preprint Let Androids Dream of Electric Sheep on arXiv, check it out!
- May 2025 Our paper CII-Bench has been accepted to ACL 2025 Main, check it out!
- Apr 2025 I'm honored to be invited to be a reviewer of NeurIPS 2025!
- Feb 2025 I'm honored to be invited to be a reviewer of ACL 2025!
- Jan 2025 It's my great honor to join VisionX Lab as a student intern, advised by Prof. Saining Xie!
- Jan 2025 Our project EmoLLM has owned 1K+ Stars!
- Jan 2025 I'm honored to be invited to be the hack lead of Spark Lab S2, let's make something big!
- Dec 2024 I will attend NeurIPS 2024 @Vancouver, please feel free to connect with me!
- Nov 2024 Our paper CPsyExam has been accepted to COLING 2025, check it out!
- Oct 2024 We preprint CII-Bench on arXiv, check it out!
- Sep 2024 Our paper II-Bench has been accepted to NeurIPS 2024 Datasets & Benchmarks, check it out!
- Aug 2024 I'm honored to be invited to be a reviewer of ICLR 2025!
- Aug 2024 I will attend ACL 2024 @Bangkok, please feel free to connect with me!
- Jul 2024 We have a short talk about CPsyCoun at AI TIME, check it out!
- Jun 2024 We preprint II-Bench on arXiv, check it out!
- Jun 2024 We have a long talk about CPsyCoun in Shanghai AI Lab, check it out!
- May 2024 Our paper CPsyCoun has been accepted to ACL 2024 Findings, check it out!
- Apr 2024 Mini-Sora has owned 1K+ Stars (my first time)!
- Mar 2024 EmoLLM, the project I lead, won the Innovation and Creativity Award in the 2024 Puyuan Large Language Model Series Challenge Competition (Spring) held by the Shanghai AI Lab!
- Mar 2024 Mini-Sora, the project I co-lead, has owned 900+ Stars and has been posted on GitHub Trendings for two days continuously!
- Mar 2024 We preprint GraphInstruct on arXiv, check it out!
📝Publications & Preprints
-
Blog
Open-source project, 2026. -
Multimodal Reinforcement Learning Team @ Tencent HY, 2026.PDF Core Contributor
-
arxiv
arXiv preprint, 2026. -
CVPR'26 CogVL
Cognitive Foundations for Multimodal Models Workshop at CVPR, 2026. -
ACL'25
The Annual Meeting of the Association for Computational Linguistics (ACL), 2025. -
NeurIPS'24
The Annual Conference on Neural Information Processing Systems (NeurIPS) Datasets and Benchmarks Track, 2024. -
ACL'24
The Annual Meeting of the Association for Computational Linguistics (ACL) Findings, 2024. -
COLING'25
The International Conference on Computational Linguistics (COLING), 2025. -
FCS'25
Frontiers of Computer Science (FCS), 2025. (CCF-T1)
⚔️Projects
- Leader of EmoLLM
, large language models for mental health. Built the end-to-end framework spanning data construction, post-training, evaluation and deployment. Post-trained 20+ LLMs (7B–32B) with SFT, OPD, and RL (DPO, PPO, GRPO, etc.), and scaled it to one of China’s largest LLM-for-mental-health communities (15k+ users).
- Co-Leader of Mini-Sora
, a community aims to explore the implementation path and future development direction of Sora.
- Core Contributor of OpenDILab
, the first decision intelligence platform covering the most comprehensive RL algorithms and applications both in academia and industry. Contributed to LightRFT — RLHF/RLVR fine-tuning for LLMs and VLMs over vLLM and SGLang backends with FSDP2/ZeRO sharding — and to the core frameworks DI-engine and LightZero.
- Core Contributor of Awesome LLM Strawberry
, a collection of LLM papers, blogs, and projects, with a focus on OpenAI o1 and reasoning techniques. Led the literature review on LLM reasoning and thinking with images.
- Core Contributor of Awesome RLHF
, a collection of research papers for Reinforcement Learning with Human Feedback (RLHF), one of four curated surveys I maintain on RLHF, RLVR, UI/GUI agents and exploration in RL.
- Core Contributor of Tianji
, a project about large language models (LLMs) and human social skills.
- Contributor of InternLM
, a community of high-quality LLMs as well as a full-stack toolchain for development and application.
- Contributor of verl
, a flexible, efficient and production-ready RL training library for large language models (LLMs).
- Contributor of OpenMMLab-mmagic
, an open-source toolbox about image and video editing&generating based on PyTorch.
💼Education & Experience
Education
Huazhong University of Science and Technology
Wuhan, China
Research Experience
Tencent Hunyuan
Supervised by Tianyu Pang
Focus: Multimodal RL post-training, Long-horizon visual agent, Multimodal evaluation, Video generation.
VisionX Lab, NYU Courant
Supervised by Prof. Saining Xie
Focus: Thinking with images, Spatial intelligence, Explicit visual states, Visual state tracking and memory, Multi-step visual inference.
OpenDILab, Shanghai AI Lab
Supervised by Yazhe Niu, Chaochao Lu and Prof. Hongsheng Li
Focus: Multimodal RL post-training, Agentic RL for long-horizon trajectories, Latent world models, Open-source RL infrastructure.
High Performance Data Mining Lab, CAS-SIAT
Supervised by Minghuan Tan, Prof. Shiwen Ni and Prof. Min Yang
Focus: Multimodal benchmarks, Image implication understanding, Human-centered evaluation of language models.
Cluster and Grid Computing Lab, HUST
Supervised by Prof. Hong Huang and Prof. Hai Jin
Focus: Graph understanding and reasoning, Instruction tuning for large language models.
🏆Honors & Awards
- 2025 SenseTime Scholarship, SenseTime (Top 30 undergraduates nationwide, <0.01%)
- 2025 Baosteel Scholarship, Baosteel Education Foundation (Top 0.1% nationwide, 509 awardees)
- 2024 China National Scholarship, Top Prize, Ministry of Education, China (Top 60 awardees nationwide)
- 2024 NeurIPS Scholar Award
- 2024 Innovation and Creativity Award, Shanghai AI Lab & GDC (Top 5/122)
- 2023 2nd Place, iFLYTEK AI Developer Competition (2/2682)
- 2023 Best Creativity Award, ZhenFund & SegmentFault & Monica Hackathon (1/80+)
- 2023 Global Rank 6th, Founder Park & Zilliz & ZhipuAI Global Hackathon (6/240+)
💡Talks & Reports
- 2026青源学术年会成功举办，青年力量共探“走向世界的模型” - Interview by BAAI
- 2025商汤奖学金公布！30位“AI新星”亮相 - Interview by SenseTime
- ACL 2025 Talk：Can MLLMs Understand the Deep Implication Behind Chinese Images? - Interview by AI TIME
- ACL 2024 Talk：CPsyCoun - Interview by AI TIME
- 当大模型遇上心理咨询：基于心理咨询报告的对话重构和评测框架CPsyCoun - Interview by Shanghai AI Lab
- 大模型创意无限：异彩纷呈的GLMs个性化智能体带来全新体验 - Interview by Zhipu.ai
🎓Services
Conference Reviewer
The Annual Conference on Neural Information Processing Systems (NeurIPS) 2025 & 2026 The International Conference on Learning Representations (ICLR) 2025 & 2026 The International Conference on Machine Learning (ICML) 2026 The Annual Meeting of the Association for Computational Linguistics (ACL) 2024 - 2026 The Conference on Empirical Methods in Natural Language Processing (EMNLP) 2024 - 2026 The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2026 The European Conference on Computer Vision (ECCV) 2026
Community
Shanghai AI Lab & InternLM Camp Teaching Assistant (TA) BAAI Qinyuan Member M-A-P Research Group Member HuggingFace China Member OpenMMLab Member DataWhale Member AI TIME Member SparkLab S2 HackLead