• Designed and implemented RESTful APIs to enable ad content management, billing configurations, customer verification, and permission control, supporting enterprise users in managing millions of ad campaigns, pages, and account users at high transaction frequency.• Led a team in developing distributed financial data processing workflows, ensuring billing accuracy, timeliness, and resilience to data inconsistency and loss.• Developed a machine learning-based customer credit score calculation pipeline using XGBoost and integrated prediction scores into ad serving systems, optimizing spending and preventing fraud.

Used Java Spring and C# ASP.NET to develop the acquisition management platform that processes leases, purchases, and trade-ins at Tesla.com. Improved the stability and response time of the used car valuation service.

Full-stack software engineer at Teaming Platform. ⚬ Developed a web-based tool that generates analysis of student survey data using Python, JavaScript, and Google Cloud Platform (GCP). ⚬ Designed and implemented a NoSQL database for storing student and class information using Cloud Firestore. ⚬ Designed and implemented the Cloud Functions that read Excel files containing student information and survey results and store them in Firestore using JavaScript. ⚬ Implemented a data pipeline for generating survey analysis using Python and deployed it as a Google Cloud Service. ⚬ Designed and implemented webpages for uploading surveys and invoking the other components using Node.Js and deployed it on Google Cloud.

Research and Develop Intern at Intelligent Cloud and Edge Group, Microsoft Research Asia (MSRA).Designed and implemented a document match plan generation algorithm for the Bing search engine based on deep reinforcement learning, improving the match plan’s performance by 57.03%.

Implemented an integrated interactive traffic visualization tool based on Uber AVS.Implemented an imitation learning model for autonomous driving using PyTorch and OpenCV.Proposed a new imitation learning model for trajectory prediction based on CNN.

Implemented a deep learning model for human pose estimation using TensorFlow.Designed an integrated system for multimodal emotion recognition. Proposal approved by Google Summer of Code 2019.

Supervised by Professor Alexandre Bayen.Joint project with Uber on mobility modeling.Tested a transportation optimization system.Visualized and analyzed data on transportation statistics using Python.

Supervised by Professor Tao Ren.Earned second prize in Chinese Artificial Intelligence Design Contest.Designed and tested an interactive humanoid robot based on Android and Microsoft Kinect.Proposed an object searching method based on computer vision and collaborative filtering.

Built a development environment Implemented a RESTful API for smartwatches in Java using Docker, Jenkins, Nginx, and Nexus3 on a Linux server.