• Designed and developed a Asset Management System to help Meta conduct automatic transaction processing, data analysis and reporting for its $50 billion in cash and $133 billion in total assets.• Designed an intuitive, scalable and secure front-end interface for the Asset Management platform with React.js and Relay, and built the data query, mutation and creation operation using GraphQL. Integrated tasks that previously needed to be done in multiple systems into one system, reducing operation times by 70% and automating 90% of the transaction work.• Developed a fault-tolerant back end service using Hack/PHP and designed a privacy-aware and object-oriented data structure with Ent framework that reduced data load times by 87% (2s down to 0.3s) compared to using raw SQL.• Simplified exception handling by providing actionable feedback to users by assigning jobs to Fusion and pulling data from Oracle asynchronously, saving 8+ developer hours per person/week and removing friction in workflows.• Refactored and deployed Meta’s Role-Based Access Control mechanism to exception handling system by extending its Ent Framework and GraphQL, securing internal system log files in Meta’s core system.• Re-architected the query system from the root level, providing better scalability, modularity for hundreds of projects.• Improved the code quality by designing end-to-end, unit and integration tests with Jest testing framework to cover the entire workflow, increasing the code coverage from 55% previously to now 98%.

• Built a real-time video sharing website with JavaScript, React.js, Grommet, and RESTful APIs that was used by 6+ million people in China; Designed and developed the core real-time video sharing engine that processed and handled 30+ million videos per day.• Designed and implemented an intelligent downloading solution that consisted of a new download module and resumed downloads by using the OkHttp framework and switching the download status via ConnectivityManager in Android. Launched A/B testing which found a 20% download efficiency increase for users.• Refactored the video transfer stream module by utilizing libVLC in Java which reduced the average live streaming latency by 80% from 1 second to 0.2 seconds.• Refactored a Java thread pool to better support async tasks and high concurrency access, reducing delays by 30%.• Designed a service to automatically send system logs to the Scuba database warehouse for further research and analysis. Deployed the service in the AWS Elastic Container Service with a CI/CD pipeline for automation.