- Ads Privacy AI: on-device predictions, federated learning- Real-time AI

I spent my entire 4+ years at Amazon as a Principal Engineer working on Alexa AI. Some notable initiatives that I helped deliver included:- Customer satisfaction detection service- ML-based ground truth generation- Affinity-based personalization- ML-based customer satisfaction metric, computed in near real time- Cross-skill recommendation (a skill is like a mobile app in the Alexa world)- A/B adoption improvement

- Machine learning insights for automated bidding in DoubleClick Bid Manager- Google Chat Smart Reply- Google Inbox- Google Compare Auto Insurance- Gmail backend- YouTube backend

Worked on an NSA-sponsored joint project between Rockwell Collins and the University of Minnesota to build a verification system for specifying, implementing, and certifying high assurance network guards. Led the design and implementation of a verification model to check recursively defined data structures. Co-inventor of the Guardol domain-specific language for guard (firewall) applications.

Led a team of 10 engineers to develop a multimedia search engine from scratch. Launched as part of a default mobile homepage for several Vietnamese biggest telcos.

Led the design and development of a static analysis system for the set/bag constraints, which allowed users to verify 98% of functions in the OCaml List library.

Designed and implemented a test case generator for the floating-to-fixed-point conversion using metaheuristic algorithms. Designed and implemented a linear-time algorithm to statically compute the heap memory used in Java Card applets. Taught college-level Computer Science classes.