LinkedIn Identity and Profile platform services.- Worked on LinkedIn Identity and Profile platform services for 600+ million members handling over 500,000 queries per second- Merged backend systems to reduce the Profile service by 7% query fetching latency and 28% memory usage- Onboarded 600+ million LinkedIn members to new Profile Bing Geo location service- Handling Profile read/write requests to LinkedIn unified content filtering service to provide spam & bad-content filtering- Improved LinkedIn public url services to resolve race condition and database sync issues- Developed and published /memberRichContent endpoint on LinkedIn Restli Gateway to serving vectorized content on LinkedIn Profile- Onboarded vectorized content service to 600+ million members - Developed spark scripts to check and maintain Database integrity- Merged captain metadata to Profile codebase to onboard Profile services with LinkedIn Anti-Abuse services- Developed new Page-View service pipeline for SlideShare- Onboarded all Page-View events to the new pipeline for SlideShare

Leading team to build/maintain central flow control system and provisioning scripts of Oracle Cloud Infrastructure with REST/SOAP APIs by Java/Perl/Python. Building testing automation and monitoring interfaces to track systems of Oracle Cloud behavior/performance. Dealing with bugs/issues timely once they were discovered.Proposing new features or sending feedback to the bugs filed from external/internal customers in Oracle daily meetings. Speaking for team achievements and negotiating for project schedules.Tracking release schedule with cross-team leaders from front-end, shopping portal, service consoles, and representatives from CEO/VP offices to meet production criteria.

Built sophisticated 3-D scanner for specular objects using low-cost depth sensing camera, and reconstructed noise-polluted signals by energy optimization theory, reducing at least 30% mean squared error than other related works.

Constructed network structure of the command for general and maneuvering usage respectively, protected information security of the command, as well as maintained computers, networks, and communication facilities.

Built relational term suggesting system by conducting social network analysis techniques and soft clustering theory to hierarchical terms-to-contributors model in Wikipedia, reporting synonyms, antonyms, and logically related terms simultaneously which other related works were unable to do, and increasing 5% precision on average than other works.