Ads Platform, AI infra, Data Platform

• Designed and implemented a new proto data storage for GPay application, which allows for multi-process accessing with the same data consistency, reduceing the read performance from 4ms to 1ms for 16kb data set.• Built the new end to end P2P pay flow, which includes implementing the new client UI based on Flutter framework, designing and creating new fast, stable and secure money movement API for the backend server.• Designed and implemented entire Cache waitlist signup flow. Built the new client services and pages, designed and implemented the new database storage schema and developed two new RPC services which have ~0.1 and ~0.9 QPS. This flow has already helped GPay Cache grow over 100K new users.• Led GPay device binding project, which enables strongly-authenticated transactions that are low friction to the users during the purchase/signup flow. Designed and developed the new devcice binding flow from scratch. Built the fast, fault-tolerlant and secure device authentication mechanism based on the FIDO protocols.

• Refactored the whole system(JAM community) into modularized single page application with Ember.js and Ruby on Rails.• Joined the development of automation test system, and used Watir to build the system’s framework.

• Implemented and verified the new algorithm for SyGus problem based on SMT solver.• Optimized the algorithm by applying different model, eliminating redundant calculation and shortening critical path, which reduced its average running time from ~100s to ~0.5s without degrading accuracy.