Software Development Engineer Intern at Amazon

• Applied Apple’s iOS Edge Saliency Model within the Vision framework supported by Neural Engine to generate bounding boxes for salient regions and performed smart cropping on images in Swift.
• Assessed model performance on more than 1.5k images across diverse image datasets, including selfies, food, and landscapes.
• Developed a visual prototype in XCode showcasing the real-time impact of smart cropping and enabled side-by-side comparison with the baseline model. Work submitted to ECCV2018.
• Conducted comparative analysis among the iOS Edge Saliency Model, the baseline model, and the Amazon In-House Saliency Model. Evaluated performance metrics such as latency and IoU score, providing actionable insights and suggestions for future enhancements.
• Uploaded the test beta version to Testflight to ensure the functionality’s accessibility in later builds.

Research Assistant at UNC Graphics & Virtual Reality Group

• Trained a ResNet body-segmentation model on human body images to perform human body segmentation task, effectively preprocessing the data and providing as an prior to the 3D body reconstruction project.
• Creatively harnessed the temporal frames information by integrating propagation module and optical flow analysis, resulting in significant enhancement of video segmentation accuracy and the overall IoU score.

Backend Intern at Shenzhen Gridsum Technology Co., Ltd.

• Integrated 20+ APIs utilizing Java Spring Boot, MyBatis, MySQL, Postman and API documentations, greatly expanding app functionalities.
• Wrote comprehensive test suites leveraging Mockito, achieving test coverage exceeding 80% across features.

Research Assistant at Biomedical Research Imaging Center IDEA Lab at UNC Chapel Hill

• Led development and training of Convolutional Neural Networks (CNN) on Alzheimer’s Disease Neuroimaging Initiative (ADNI) datasets, resulting in a three-way classification model with great accuracy.
• Utilized transfer learning and attention mechanisms to identify key Alzheimer’s Indicator brain region.

Instructional Assistant at Department of Statistics and Operations Research at UNC Chapel Hill

• Graded assignments for a “Data Analysis” course for over 200+ students using R and provided feedback to help students better understand the coursework.