About Me

I am currently a Ph.D. student in Computer Science and Engineering at HKUST, advised by Prof. Yi R. Fung. My research focuses on multimodal large language models, web agents, multimodal reasoning, and deep research.

I am interested in building agent systems that can search, browse, reason, use tools, and learn from long-horizon trajectories. My recent work studies multimodal deep research agents, trajectory generation, reflection, self-distillation, RLVR, and large-scale SFT data for tool-use alignment.

News

Jun 2026 GeoBrowse accepted to ECCV 2026.
Jun 2026 Joined Zhipu AutoGLM, working on AI for AI.
2026 WebWatcher accepted to ICLR 2026 as a poster.
2025 Contributed to the Tongyi DeepResearch Technical Report.
Nov 2025 Joined Tencent AI Lab / Hunyuan Frontier Group.
Sep 2025 Started Ph.D. study at HKUST.
2025 ParseCaps accepted to AAAI 2025 as a poster.
2024 OrthCaps accepted to CVPR 2024 as a poster.

Education

The Hong Kong University of Science and Technology
Ph.D. in Computer Science and Engineering
Sep 2025 - Present
Harbin Institute of Technology, Shenzhen
M.Eng. in Electronic Information · GPA 3.365/4.0 · Rank 12/82
Sep 2022 - Jun 2025
University of Electronic Science and Technology of China
B.Eng. in Automation · GPA 3.61/4.0
Sep 2018 - Jun 2022

Internship

Zhipu AutoGLM
Jun 2026 - Present
Working on AI for AI, with a focus on agentic systems and automated AI workflows.
Tencent AI Lab / Hunyuan Frontier Group
Nov 2025 - May 2026 · Shenzhen, China
Working on deep research agents, multimodal web agents, trajectory reflection, on-policy self-distillation, and RLVR for agent learning.
Alibaba Tongyi Lab
Jul 2024 - Aug 2025 · Hangzhou, China
Built multimodal deep research agent systems, tool-use trajectory generation, search/browse/reasoning pipelines, and large-scale SFT data for tool-use alignment.

Selected Publications

Multimodal Agents & Deep Research
WebWatcher: Breaking New Frontier of Vision-Language Deep Research Agent
X. Geng*, P. Xia*, Z. Zhang*, et al.
ICLR 2026 Poster
Search, Distill, Evolve: Self-Distillation of Web Agents in Verifiable Virtual-Tool Environments
X. Geng, T. Fang, Z. Liang, et al.
NeurIPS 2026 Submitted
GeoBrowse: A Geolocation Benchmark for Agentic Tool Use with Expert-Annotated Reasoning Traces
X. Geng*, Y. Xiao*, et al.
ECCV 2026
Efficient Multimodal Planning Agent for Visual Question Answering
Z. Chen*, X. Geng*, X. Wang, et al.
ICML 2026 Under Review
Tongyi DeepResearch Technical Report
TDR Team
Technical Report 2025
Vision-Language Models & Medical AI
Detecting Knowledge Boundary of Vision Large Language Models by Sampling-Based Inference
Z. Chen, X. Wang, Y. Jiang, Z. Zhang, X. Geng, et al.
EMNLP 2025 Main
OrthCaps: An Orthogonal CapsNet with Sparse Attention Routing and Pruning
X. Geng, J. Wang, J. Gong, et al.
CVPR 2024 Poster
ParseCaps: An Interpretable Parsing Capsule Network for Medical Image Classification
X. Geng, J. Wang, J. Xu, et al.
AAAI 2025 Poster
AgentVista: Evaluating Multimodal Agents in Ultra-Challenging Realistic Visual Scenarios
Z. Su, J. Gao, H. Guo, Z. Liu, L. Zhang, X. Geng, et al.
ICML 2026 Under Review
Struct-Searcher: Agentic Structural Thinking Advances Multimodal Deep Information Seeking
F. Zhang, G. Zhang, S. Qian, H. Li, X. Geng, et al.
ICML 2026 Under Review

Academic Service

Reviewer for ICLR, ICML, ECCV, AAAI, and CVPR.

Invited talks at MLNLP, Beijing Academy of Artificial Intelligence Forum, and AAAI Imageomics Workshop.

Awards

  • National Scholarship, 2024.
  • Special Academic Scholarship, Harbin Institute of Technology, 2024.
  • Third-Class Entrance Scholarship, Harbin Institute of Technology, 2022.
  • University Scholarship, UESTC, 2018-2021.