Chen Fuhao

XPeng

• Responsible for data-flow architecture design, data quality and cost optimization of autonomous driving data of various vehicle models
• Responsible and established the very first image and video data cloud desensitization service within the company and applied to domestic and overseas privacy protection
• As a tech lead to deliver projects, coordinate team resources and reduce technical debts

Grab

• Constructed a geographic traffic flow platform covering 100+ cities in Southeast Asia, and formulated the task scheduling workflow and the entire CI process
• Refactored the real-time traffic computing system, modularized and separated each component, and realized the whole CD process based on k8s
• Real-time vehicle GPS aggregation and watermark calculation are realized, which improves the accuracy of ETA and the reasonableness of billing
• Coupling offline data into real-time calculations through source modularization, reducing system complexity
• Migrating the traffic platform from EMR to EKS, improving machine utilization and application stability

Tencent

• Introduced the hierarchical split of automated sharding and indexing, which increased the log indexing speed of the game log retrieval platform by 30%
• Designed a set of unified interfaces to realize the routing, migration, hierarchization, expansion and contraction, monitoring, and failure recovery of data in different database components
• Constructed a set of multi-database component management system to improve the efficiency of database management automation operation and maintenance
• Open source a relatively independent, complete es automated monitoring dashboard

JD.com

• Designed an advertising smart bidding platform, which increased 5% ROI of the middle and external pages of the advertising platform
• A multi-dimensional valuation system has been implemented to help advertisers achieve automatic differentiated bidding in different channels and refine their advertising budget allocation
• Constructed a set of product retrieval system for WeChat social users, reached more than 10 millions of GMV per month and attracted millions of brand new users
• Introduced the multi-touch attribution(MTA) model and confidence interval to solve the problem of sparse advertising data in different dimensions