Build a cross-screen device association solution based on data at petabyte level on Hadoop/Hive/Pig/Apache Spark environment. Build the data aggregation/ETL process for ~3 billion records per day. The association solution covers hundreds of millions of devices in the US.Apply and implement varies machine learning and data mining approaches to classify the key dimension of the association solution, as well as improve the accuracy of the association decision, significantly increased the business value of the product.Build a near real-time enterprise analytic dashboard that monitors job status, delivers warning messages, and displays hundreds of analytical data points from different dimensions, therefore providing a single interface for all insights needed for the overall data platform. The dashboard is powered by fast data analytics based on Spark and Google BigQuery.

Designed and led the implementation of a big data cloud service for Belkin’s new Echo product line. Built a Cassandra-powered cloud application that’s capable of handling extreme large amount of sensor data from hundreds of thousands of devices. Worked in the highly scalable Linksys Smart Wi-Fi cloud development project, designed and implemented Restful APIs for the Linksys Smart Wi-Fi Cloud which handles millions of Linksys smart routers' daily activities.

Designed and implemented the center piece of a data warehouse for one of the nation’s largest Starter repository– the ETL application that imports the data. The system was designed with Service-Orientated Architecture and is highly configurable to accommodate numerous different source systems, as well as constant changes on the business rules of the import process. The light weighted J2EE system utilizes Hibernate framework, and runs on an IBM WebSphere J2EE server cluster. The highly scalable and configurable design minimized the maintenance cost and down time while successfully accommodated the constant changes of business requirements, making the repository one of the most important data source in the business.Design and implemented a Web Service interface for a data warehouse with more than 20 million records. As the lead developer, worked closely with the DBA team on designing and implementing the data warehouse on Oracle 10g/11i. The web service interface was a success to fulfill their business needs and is able to return search results in the matter of a couple of seconds.As development team lead, helped to build a second-level technical support team for the systems described above. Designed and implemented a nationwide document delivery and distribution sub-system for the business’ major production system. The queue based, highly distributed system delivers tens of thousands generated documents to offices across the country on time to fulfill business need.