About Me
I am a Senior Algorithm Engineer at Xiaomi Robotics Lab, focusing on manipulation and loco‑manipulation using Vision-Language Models (VLMs) and Vision-Language-Action Models (VLAs). I am seeking self‑motivated interns to work with me on agentic embodied AI.
I received my Ph.D. from the Institute of Software, Chinese Academy of Sciences. My research interests include real‑time systems, embedded AI, and reinforcement learning. I was fortunate to be advised by Prof. Ying Qiao at the Human–Computer Interaction Technology and Intelligent Information Processing Laboratory.
I previously worked with Prof. Junliang Xing at the Institute of Automation, Chinese Academy of Sciences. We placed third and fourth in the CIG 2017 and AIIDE 2018 StarCraft AI competitions. I am a Grandmaster‑level StarCraft player and passionate about applying AI to games.
Email: bozitong@xiaomi.com, bozitong1996@gmail.com
I share study notes on Zhihu from time to time—feel free to follow. Zhihu
News
- Our paper Reinforced Embodied Planning with Verifiable Reward for Real-World Robotic Manipulation is now on arXiv.
- I will visit NUS and NTU on September 3–4 to introduce Xiaomi Robotics Lab. We welcome like‑minded students to join our team.
- I officially joined Xiaomi Robotics Lab on January 21, 2025.
- Our work RTDeepEnsemble: Real-time DNN Ensemble Method for Machine Perception Systems has been accepted to the 42nd IEEE International Conference on Computer Design (ICCD 2024); I will give an oral presentation on November 18 in Milan, Italy.
- Our latest work on energy management for embedded systems using reinforcement learning has been submitted to the 22nd ACM Conference on Embedded Networked Sensor Systems (SenSys 2024).
- Our papers Designing Real-Time Neural Networks by Efficient Neural Architecture Search and HFGCN: Hybrid Filter Graph Convolutional Network for Heterophilic Graphs have been accepted by the 2024 International Conference on Intelligent Computing (ICIC 2024). See you in Tianjin, China!
- Our work Developing Real-Time Scheduling Policy by Deep Reinforcement Learning was accepted by the 27th IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS 2021). I will give an oral presentation.
Visitor Map