Zhaoxin Meng

I am currently a Staff Machine Learning Engineer at HONOR Robotics Lab, focusing on Embodied AI, Vision-Language-Action (VLA) models, and World Models.

Previously, I worked as a Senior Machine Learning Engineer at DiDi Global (Voyager Group), specializing in autonomous driving perception and generative data synthesis. I received my Master’s degree from Beijing University of Posts and Telecommunications (BUPT) in 2022.

My research interests lie in Robotics Perception, Large Multimodal Models, and Data Synthesis for autonomous systems.

💼 Work Experience

Internships

• DiDi Global: Autonomous Driving Perception (2021 - 2022)
• ByteDance (TikTok Ads): Deep Learning & CVR Prediction (2021)
• Microsoft Research Asia (MSRA): Multi-Modal Document Understanding (2020 - 2021)
• SenseTime: ADAS Perception (2020)
• ByteDance (AI Lab): Speech & Deep Learning (2019 - 2020)

🎓 Education

• Beijing University of Posts and Telecommunications (BUPT)

  M.E. in Computer Science (2019 - 2022)

• Donghua University

  B.E. in Network Engineering (2015 - 2019)

• Mid Sweden University

  Exchange Student (2018 - 2019)

📝 Selected Publications & Patents

Papers

• [IJCNN 2021] Meng Zhaoxin, et al. MF-NET: Meta fusion network for 3d object detection.
• [ITAIC 2019] Wang Shaoyu, …, Meng Zhaoxin, et al. Edge-aware fully convolutional network with CRF-RNN layer for hippocampus segmentation.

Patents

• [CN117372997A] An obstacle recognition method, device, vehicle and electronic equipment. (2023)
• [CN114332494A] 3D object detection and recognition method based on multi-source fusion in vehicle-road collaborative scenarios. (2021)

🏆 Honors & Awards

• Excellent New Graduate Award, DiDi Voyager (2023)
• Star of Tomorrow, Microsoft Research Asia (2021)
• Bronze Award, ACM-ICPC Beijing Regional (2017)