Zhang Dongshuo (张东烁)
I am currently an AI Researcher in Embodied Robotics at Sony. Prior to Sony, I worked at the Hyundai-NTU-A*STAR Corporate Lab with Prof. Kong Wai-Kin Adams, and lead the SLAM research group at the HESL, CCDS, NTU under Prof. Lam Siew Kei. My research focuses on AI for Robotics, with a particular interest in Vision-Language-Action (VLA) models, Robot Localization and Perception, Visual Simultaneous Localization and Mapping (SLAM), and Visual Place Recognition (VPR).
I earned my Ph.D. from CCDS, NTU, Singapore in Jan 2026, under the supervision of Prof. Lam Siew Kei, and my M.S. in Electronic Information Engineering (EIE) from City University of Hong Kong (CityUHK) in Oct 2019, under the supervision of Prof. So Hing Cheung (IEEE Fellow).
Previously, I worked at CamSense (欢创科技) as an algorithm engineer from 2017 to 2018, and at RoboMaster, DJI (大疆创新) as an embedded software engineer from 2016 to 2017.
News
- [2026-03] Farewell to my five years at NTU. I’m excited to start a new journey at Sony as an AI Researcher in Embodied Robotics.
- [2026-01] I’ve passed my PhD Oral Defence.
- [2025-09] I’ve started a new role as a Research Associate @ Hyundai-NTU-A*STAR Corporate Lab.
- [2025-06] 🎉 Congrats Gao Ye! Our work on Motion-Feat: Motion Blur-Aware Local Feature Description for Image Matching has been accepted to IROS 2025.
- [2025-04] 🎉 Congrats Nanhua! Our work on SHAA: Spatial Hybrid Attention Network with Adaptive Cross-Entropy Loss Function for UAV-view Geo-localization has been accepted to IEEE T-CSVT.
- [2025-01] I’ve started part-time teaching the undergraduate course SC3000/CZ3005 Artificial Intelligence at CCDS, NTU.
- [2024-06] 🎉 Our work on CurricularVPR: Curricular Contrastive Loss for Visual Place Recognition has been accepted to IROS 2024.
- [2023-06] 🎉 Our work on Training-Free Attentive-Patch Selection for Visual Place Recognition has been accepted to IROS 2023.