Chief Scientist | Managing Director of Research, Face++ (Megvii Technology Ltd.)
After thirteen amazing years at Microsoft Research, I have joined Face++ (Megvii Technology) (in July, 2016) as Chief Scientist and Managing Director of Research.
Jian was born in Xian, China, which is home of The Terracotta Army. He received a B.S., M.S., and a Ph.D. degree from Xian Jiaotong University in 1997, 2000 and 2003, respectively. Immediately following, he joined Microsoft Research Asia, and has been working in the fields of computer vision and computer graphics, with particular interests in solving fundamental research problems and building real-world working systems. His primary research interests are computational photography, face recognition, and deep learning based image understanding.
Email: sunjian AT megvii.com
Job/Internship opening: Researcher/Engineer in Beijing/Nanjing/Shanghai/Chengdu China or Seattle US.
- Megvii (Face++) team won COCO and Mapillary Challenges 2018: 1st Places of COCO Detection, COCO Panoptic Segmentation, COCO KeyPoints, Mapillary Panoptic Segmentation.
- Yichen Wei joined us as Director of Megvii Research Shanghai, China. Here is the job description.
- We pleased surprised that ShuffleNet V2 (Practical Guidelines for Efficient CNN Architecture Design) works so well on small/large model on ARM/GPU.
- Megvii (Face++) team won COCO and Places Challenges 2017: 1st Places of COCO Detection, COCO Keypoint, Places Instance Segmentation.
- ResNets was applied in DeepMind's AlphaGo Zero : a 40/80-layer ResNet for predicting AlphaGo's move selections and the winner of games.
- Our ShuffleNet (An Extremely Efficient Convolutional Neural Network for Mobile Devices) paper is available.
- Jue Wang joined us as Director of Megvii Research US, a new AI lab based in Seattle. Please contact Jue for job or intern opportunities.
- Established Megvii Research, a R&D department for both fundamental and applied research, focusing on deep learning algorithm, infrastructure, and data science.
- AI on Cloud (Microsoft Azure's Cognitive Services), my team contributed Computer Vision APIs and Face APIs.
- Microsoft's CaptionBot, empowered by ResNets.
- Won both ImageNet and MS COCO competitions 2015, by 152-layer neural networks (ResNets) . (technical paper and press stories 1, 2, 3, 4, 5, 6, ...).
- How-old.Net, a phenomenal cloud service, hundreds of millions of visits in 2015. The magic behind was from my team member Xudong Cao, who is running Momenta.
- Windows Hello (video), we shipped on Surface Pro 4/Surface Book.
- Kinect Identity (video), a key component we contributed to Microsoft's Xbox Kinect, for recognizing and tracking player identity.
- COCO and Mapillary Challenges 2018 Competition Winner Award, 2018
- COCO and Places Challenges 2017 Competition Winner Award, 2017
- The National Natural Science Award of China (2nd Class), 2016
- Best Paper Award, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016
- ImageNet Large Scale Visual Recognition Challenge 2015 Competition Winner Award, 2015
- COCO Challenge 2015 Competition Winner Award, 2015
- MIT Technology Review Young Innovator under 35 (MIT TR35), 2010
- Best Paper Award, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2009
- Microsoft Fellowship, 2001
- 1st Place in the Graduate Student Examination, Xian Jiaotong University, 1997
- Area chair: ICCV 2011, CVPR 2012, CVPR 2015, CVPR 2016, CVPR 2017, CVPR 2019
- Program committee: Siggraph 2010