1. Worked on the Audible backend search team, supporting both customer-facing and internal applications using Java, AWS, and Kanban management tools.

2. Designed and implemented the Audible search autocomplete widget experiment on the Amazon central search bar in the U.S. and worldwide, including zero-downtime cache migration and CI/CD pipeline split for service decoupling.

3. Led the development of a serverless data validation platform for internal use, featuring a React frontend, user authentication, an AWS-based backend, and fulfillment of security requirements.

4. Engaged in project planning, process tracking, and communication as the project owner.

Generated dataset for Scientific Discovery task and proposed several baseline methods based on language model and graph neural network. As the first author, the result was published in the ICLR 2022 Deep Learning on Graphs for Natural Language Processing Workshop.

Built interactive UI for researchers to visualize our datasets and results.

Led independent research about cross-domain open-set classification in the area of computer vision. As the first author, one paper was published in the 2021 IEEE International Conference on Image Processing (ICIP) main conference.

Developed open-source platform for AI-driven cellular electron cryo-tomography analysis.