Xin Guo Email and Phone Number

Lead multiple MLE teams and ML infra team to improve Discord product user experience and increase user engagement by leveraging ML and Gen AI

I am leading the ML/data platform that is in charge of two important domains of Snap content org: The first domain is about ML infra covering feature and training data generation and serving, with a track of records for increasing system scalability, reducing signal latency from batch to real time, accelerating E2E development velocityThe second domain is about building the data infra to support business reporting, decision making, product analysis, AB testing, and data compliance. In this domain I lead a team of Data Engineers to work closely with leaders/PMs/Engineers to build the the E2E ETL system with target to improve data quality, data usability, and data visualization.

Lead the team to build the core logging/data platform for Google Map with a serials of successful launches, including semantic logging, A/B test experiment tooling, and improved analysis experience for Google Map engineers. My responsibility includes goal/mission setting, project planning, project design, team cultural development, career guidance for reports, and hiring etc.I also lead another team to build the notification platform for Google Map, covering notification infrastructure and machine learning for smart targeting.

Lead the development of the centralized recommendation and ranking platform for various Instagram products, including Feed, Stories, Explore, Hashtag, Shopping. The platform has successfully served billions of Instagram users with the following advanced features:1) high reliability with low latency and efficient resource consumption2) support fast development speed by extremely easy way to do new ML experiments and launch new models3) user-friendly and comprehensive tooling for experiment, debug, and monitoring

Proposed and developed service to support third-party contacts/content ingestion into LinkedIn platform used by Google in production.Developing Hadoop offline jobs to enrich the contents of LinkedIn members' profiles.

Developed service to improve the data quality of YouTube video metadata by reconciling data from Knowledge Graph.Applied various machine learning technologies to identify incorrect YouTube video metadata.