Bid Multiplier Solution Development (Java, Protostuff/Protobuff, JSON, Docker, Yaml, HDFS)- Led a team to build bid multipliers that enable advertiser to ingest prior/posterior knowledge into the bidding pipeline. Designed the backend architecture and built the infrastructure on top of 600 distributed AdServers which dealing with 5MM request per second and response in 1ms.Ghost Bidding Solution Development (Java, Protostuff/Protobuff, Docker. Yaml, HDFS)- Led a team to build ghost bidding product feature that allows user to run A/B testing without paying for PSA ad. Integrated identity solution to enhance the A/B testing framework. Worked with 7+ teams to design the architecture and lead the end-to-end integration work and testing

Budget Allocation Enhancement (Java, SQL, Hibernate, Grafana)- Owner of budget allocation managing 2 million daily budget and designed new algorithm that applied real time spend/historical learnt delivery to facilitate the spend and reduced the underspend by 5%. F1 Console Development (Java, JS/HTML, Angular JS, Spring MVC, Hibernate)- Built end-to-end online model deployment web app that allows dynamic models deployment on distributed AdServers with reliability checking mechanism for A/B testing while providing online diagnosis and reporting.

Performance Monitoring Development on PLEX (Golang, StatsD, Graphite, Grafana)- Built poll-based monitoring application for products handling 5,000,000 rps in production.- Built automated testing framework to evaluate platform over engineering metrics- Benchmarking performance for high throughput system with golang.- Built features for performance interpretation in full stack for products.

Intelligent and Remote Health Care in Cloud Computing Environment - Created remote healthcare web service based on machine learning. - Established elastic algorithms based on machine learning for load balancing in cloud computing. Medical Product Developments in Living Lab - Designed and implemented evaluation framework for medical product development.- Developed intelligent hardware/software for monitoring daily meals allocated to patients.

Product Research and Development - Released product of concept of semantic voice assistant web service.- Launched product OPView for sentiment analysis and customer feedback analysis.- Released component for name entity annotation based on Conditional Random Field (CRF).- Released component for topic modeling.- Released component ENLP Classifier for automatic text classification & clustering.Project Development - Developed automatic text classification system for National Applied Research Laboratories.- Developed sentiment analysis system for Commerce Development Research Institute