Lei Shu

My name is 舒蕾. I am a Research Scientist at Google DeepMind, specializing in Agentic AI and Coding. I lead Project Astra 3D and am a core contributor to Project Astra and Gemini.

Prior to this, I worked at Amazon AWS AI and Amazon Alexa AI. In 2020, I earned my Ph.D. in Computer Science at the University of Illinois at Chicago under Prof. Bing Liu's supervision. I also worked as a research intern at Uber AI Lab, where I was advised by Dr. Piero Molino and Dr. Gokhan Tur. I am a winner of the Yelp Dataset Challenge.

• Agentic AI & RL — harness/loop engineering and self-evolving harnesses, multi-agent systems, agentic RL, and cross-device (server, phone, glass) agents.
• Coding — improve Gemini Vibe and Agentic Coding capability including agentic procedural 3D modeling via code.

• I am leading Project Astra 3D at Google DeepMind. Check out our latest work, 3DCodeBench, on benchmarking agentic procedural 3D modeling via code.
• At Google I/O 2026, I contributed to intelligent eyewear — watch the demo.
• I lead modeling for Computer Control & Visual Overlay on Project Astra. Demos: Action Intelligence and Education (Visual Tutor).
• Our paper Ego2Web, a web agent benchmark grounded in egocentric videos, was accepted to CVPR 2026.

• 3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code
  Yipeng Gao, Lei Shu, Genzhi Ye, Xi Xiong, Ameesh Makadia, Meiqi Guo, Laurent Itti, Jindong Chen
  arXiv 2026 [paper][project]
• Ego2Web: A Web Agent Benchmark Grounded in Egocentric Videos
  Shoubin Yu, Lei Shu, Antoine Yang, Yao Fu, Srinivas Sunkara, Maria Wang, Jindong Chen, Mohit Bansal, Boqing Gong
  CVPR 2026 [paper][project]

• Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
  Gemini Team, Google DeepMind (incl. Lei Shu)
  arXiv:2507.06261 [paper]
• Project Astra
  Project Astra Team, Google DeepMind (incl. Lei Shu)
  Google DeepMind, 2025 [project]
• UISim: An Interactive Image-Based UI Simulator for Dynamic Mobile Environments
  Jiannan Xiang, Yun Zhu, Lei Shu, Maria Wang, Lijun Yu, Gabriel Barcik, James Lyon, Srinivas Sunkara
  arXiv:2509.21733 [paper]
• Accelerating Inference of Retrieval-Augmented Generation via Sparse Context Selection
  Yun Zhu, Jia-Chen Gu, Caitlin Sikora, Ho Ko, Yinxiao Liu, Chu-Cheng Lin, Lei Shu, Liangchen Luo, Lei Meng, Bang Liu, Jindong Chen
  ICLR 2025 [paper]

• RewriteLM: An Instruction-Tuned Large Language Model for Text Rewriting
  Lei Shu, Liangchen Luo, Jayakumar Hoskere, Yun Zhu, Canoee Liu, Simon Tong, Jindong Chen, Lei Meng
  AAAI 2024 [paper]
• Fusion-Eval: Integrating Assistant Evaluators with LLMs
  Lei Shu, Nevan Wichers, Liangchen Luo, Yun Zhu, Yinxiao Liu, Jindong Chen, Lei Meng
  EMNLP 2024 Industry [paper]
• Enhancing Reinforcement Learning with Dense Rewards from Language Model Critic
  Meng Cao, Lei Shu, Lei Yu, Yun Zhu, Nevan Wichers, Yinxiao Liu, Lei Meng
  EMNLP 2024 Main [paper]