Algorithm Engineer, Ph.D.

JD.com, Beijing, China
No. 18 Kechuang 11 Street, Tongzhou District, Beijing
#C JD Building

• Led the development of JD Health's digital human technology, driving it from concept to deployment.
• Oversaw the entire project lifecycle, including technical research, dataset creation, model design, and business application.
• Developed three types of digital human generation models: a GAN-based dual-tower model, a diffusion-based end-to-end model, and a real-time model based on decoupled facial representations.
• Successfully applied these models across more than five domains, including medical knowledge dissemination, interactive dialogue, and virtual pet customization, generating over 2.5 million in annual commercial revenue and serving more than 230,000 users.
• Published two papers and open-sourced code, garnering over 1,000 stars on GitHub. [JoyVASA , JoyHallo]

• Served as a key researcher in driving JD Health’s brain science research from concept to execution, focusing on stroke rehabilitation using Brain-Computer Interface (BCI) technology and brain imaging-based diagnosis of Alzheimer’s (AD) and Bipolar Disorder (BD).
• Led the development of a proprietary EEG and brain imaging data collection platform, designed neural decoding algorithms, and managed external collaborations.
• Processed over 4,500 data samples, including private and public brain imaging datasets and more than 40 hours of motor imagery EEG data. Published five research papers and filed two patents, with multiple studies in EEG recognition and brain imaging achieving state-of-art results.
• Contributed to the successful application for a national key R&D project (Virtual Reality Cognitive Rehabilitation Training System for the Elderly) under the Ministry of Science and Technology, and helped JD Health become a member of the Brain-Computer Interface Industry Alliance.

• Research on deep learning based semantic segmentation algorithms on breast ultrasound images as well as lung MRI images.
• Dig deep into semantic segmentation algorithms, such as fully supervised learning, semi-supervised learning, neural network architecture search, etc.
• Proposed lightweight dilated densely connected network for 3D breast tumor segmentation. The performance improved over 5% compared with classical segmentation networks, while network parameters were over 20 times smaller than classical networks.
• Designed an uncertainty-aware temporal-ensembling model for semi-supervised segmentation. The performance of semi-supervised method is able to achieve 94.4% of that in supervised segmentation with only 1.1% labeled data.
• Suggested an NAS-based 3D medical image segmentation framework, and achieved an improvement of 4.2% compared with the state-of-art human-designed segmentation network.
• Related journal and conference papers have been published, total impact factor 30+.

• As a core developer, designed an automatic detection and quantitative analysis solution for surface pitting of coupling gear components in train sets, based on computer vision technology.
• Collected coupling gear surface images and applied image recognition algorithms to generate statistics, reports, and alerts for gears exceeding the defect threshold.
• Led the development of algorithms for detecting and segmenting gear surface wear areas, as well as the design and implementation of the software system on the Windows platform.
• Completed defect detection of 104 gear surfaces (internal and external) with defects larger than 1mm within 3 minutes; published related papers and patents.

• Provide solution for measuring the geometric parameters of an overhead line system using scale factors and frame differences.
• I was responsible for the previous algorithm simulation and participated in the hardware structure design work.
• Two related patents have been granted.