Experience

• 

  完美世界

  May 2007 - Jan 2010

  Software engineer

  Working responsibility: 1:Desgin and implement the general numerical system interface which has been used by company's products,including general and detail desgin,coding implemetation and testing. 2:Develop related modules of 3D engine which is a basic lower plateform for the Game's graphic design. 3:Develop the general map editor, which is suit for game's map design and creating and automatic path finding of roles.

• 

  Ericsson

  Jan 2010 - May 2012

  Data mining and network operation engineer

  Working resbonsibility: 1:Analyzing and mining the telecommunication data using various tools, and providing the timely report. 2: Providing fault prediction, detection and management on specific BSS(Business Solutions Support) network nodes: mediation, billing, CRM, data warehouse, activation & provisioning applications/servers by using data mining tools and algorithms. Perform preventive& corrective maintenance, problem management, acceptance and faults tracking. Logging and updating the performed activities in specifically ticketing tools 3: Independedly developing the small size requirement system based on LAMP enviroment. 4:Maintenance and management of Orcale database and Solaris OS behind Erisson's customer business system. Following are the project I have taken:1): Anomaly detection of network traffic of Timer Telecom Timer-Leste: Using descriptive and inferential statistics to providing low quality network situation detection for our customer.2): The whole network maintenance and operation project for Timer Telecom. Show less

• 

  华为

  May 2012 - Aug 2016

  Senior data scientist

  Lead team to provide big data solutions to famous telecommunication operators around the world. I take part in the solution's whole life circle from communicating with customers for getting requirements, to the detailed solution design, modeling, coding, testing, and finally delivery our solution to customer on time.Following are the projects I have taken:1): User’s loyalty management for China Unicom—-Using deep learning and random forest to predicting user’s complaint and churn rate to China Unicom.2): CEI(Customer Experience Index) analysis for Vodafone Spain—-Using k-means and SVD to provide real-time calculation of CEI which reflects user’s real perception and experience.3): User’s motion trajectory analysis for Etisalat UAE—-Using LSH(Locality Sensitive Hashing) to cluster user into groups who have similar motion trends, the result is as the evidence for the public transportation system design.4): High influential user recognition for TRUE Thailand—-Using community detection algorithm of social network graphs and PageRank to recognizing users who have higher social influence among others. Show less

• 

  中国平安

  Aug 2016 - now

  Senior manager - big data

  1: Lead the team to build an enterprise-level data middle platform from 0 to 1, and integrate relevant system data. In the data middle platform, offline and real-time data such as Ping An agents, customers, and products are layered and stored according to themes, thereby eliminating data islands. In terms of data governance, my team formulate data standards, improve data governance mechanisms, build data monitoring platforms and data blood relationship management platforms, and use mature mechanisms and platforms for data governance and data asset management. In terms of service applications, we develop upper-level multi-dimensional analysis engines, self-service analysis engines, data insight engines and other applications to assist business decision-making.2: Based on the ability of the data middle platform,I lead the team to build the upper-level prediction engine from 0 to 1. Firstly, the agent behavior data sequence is embeded, and deep learning modules such as CNN, LSTM, transformer and attention mechanism are used to build a deep neural network prediction engine based on the agent's full-cycle behavior. Through the prediction engine, it has successfully achieved accurate forecasting agent maintenance, performance level, and the company's NBEV (that is, profit), manpower, capacity and agent income and other indicators. And based on the forecast results, conduct root cause analysis and provide business recommendations from the data aspect. Show less