I am an algorithm strategist at Meituan. My current focus is on lightweight VLMs, model compression and inference. I received BSc from Jilin University (2010) and MSc (Laurea Magistrale) from University of Trento (2013). My early academic efforts were centered on AutoML, while I was in a NAS team at Xiaomi AI Lab.
Throughout my years in Meituan, I have led or collaborated on multiple optimization-driven projects in self-driving (2d/3d object detection, lidar auto-labelling), MLOps (covering classification/detection/segmentation/OCR algorithms for content moderation, text recognition, drones, cycling service etc.). I also built conversational systems for digital saleperson and custom service. Recently, I have been dedicated to democratizing MLLMOps within the enterprise, promoting the widespread adoption of cutting-edge AI technologies in real-world applications.
I was generously invited by Glenn Jocher (CEO of Ultralytics) to introduce our most popular open-source project YOLOv6 on YOLO Vision 2022 and its follow-up in 2023. I was once commissioned to author a tour book (2018) about Italy.
You may reach me at zhangboyd@gmail.com.