Azure Edge & Platform team is responsible for providing middle-tier technical solutions for edge computing customers and other Azure cloud teams. The team is distributed in Shanghai and Seattle.Tech stack: C#/Kubernetes/Service Fabric/Time Series Database/Prometheus/LLM- Led the development of next-gen monitoring solution for multiple Azure products, receiving positive feedback from partners and customers - Backward-compatible: supports both Microsoft's internal data format and open-source data format - High performance: memory and CPU footprint is only 30% to 50% of competitive open-source solutions (Prometheus/Cortex/Thanos) - Cross-platform and easy to deploy: support for deployment on any K8s and to both Windows & Linux nodes- a HA runtime framework that allows many of Microsoft's legacy first-party software to run directly on the latest Kubernetes platform without code modifications. This reduces the migration cost of product team from dozens of man-years to several man-months, thus accelerating the org's edge computing strategy.- Innovation project: using LLM to help Azure's customers better understand and manage their infrastructure, reducing customer's operation costs and Azure's costs on customer service.

Guangshi focuses on helping small and medium-sized company w/ their finance and tax, saving their labor costs, and assisting platform-based enterprises connecting to their merchants. Its customers include Ctrip, Quhuhu, etc.Tech stack: TypeScript/Vue/Java/Spring Boot/C#- Built and managed a R&D team of over 30 people from scratch, including frontend, backend, and Windows clients. Responsible for talent recruitment and training, planning team goals, and key technology selection and long-term technical planning.- Led the development of the first version of the system.- Data-driven R&D process: improved the R&D process to ensure data-driven and keep optimizing business and technical KPIs. I created a "Data watcher" virtual team (ref: "night watcher" in ), where each team member is responsible for different data points. I also led regular reviews, sets optimization goals and action plans, and reports progress to the entire company. As a result of the "Data watcher", the conversion rate of merchant onboarding increased by more than 10%, and the up time of our key background services increased from 99.95% to 99.995%.

It's a combination of entertainment and e-commerce, we created and operated a new type of online game: users can remotely control real claw machines through live broadcast and hardware control technology, upon successful catch, we will send them real prizes.Tech stack: objective-c/java/typescript/react/falsify/c/mqttBy the end of 2017, we achieved 15,000 new users and 3 million RMB revenue per month, and completed the angel round funding.- Designed the entire gameplay and UX from scratch.- Implemented the entire system with limited team members, including real-time broadcasting, machine control, and a self-built probability algorithm and warehouse inventory management system.- Managed the operation of the game's e-commerce system, from team building, product selection, inventory management, interactive gameplay probability, to delivery and customer service.- Managed the advertising plans, and the optimization based on ROI.

"same" app is a social network for youth.

Main program of a Netease mobile game.

Vida is a social network for posting videos & photos.

Participated in development of multiple features for Windows Server.