Securing AI Agents that Handle Sensitive Data
Industry work on threat modeling AI-agent integrations: indirect prompt injection, tool abuse, and sensitive-data exfiltration in regulated environments.
Research & Projects
Systems I have built or contributed to, from DARPA-funded research to open industry standards.
Detecting Insider Misuse of Sensitive Data
Industry work on detecting misuse of sensitive data by people with legitimate access, at the scale of regulated healthcare systems.
Privacy-Preserving Systems for Public Health and Disaster Response
QUEST (pandemic monitoring over existing Wi-Fi, with individual privacy protected) and CareDEX (NSF-funded data exchange for senior-care communities during disasters).
IoT Notary and Canopy: Verifiable Privacy for Smart Spaces
Best Paper Award (IEEE NCA 2019). Cryptographic techniques that make privacy protection in connected environments provable rather than merely promised.
TIPPERS: Privacy-Preserving IoT for Smart Spaces
DARPA-funded platform that lets smart buildings use sensor data while enforcing privacy rules for the people inside. Deployed on the UC Irvine campus and tested in the U.S. Navy's Trident Warrior 2019 and 2020 exercises.
Edge Computing and Resilient IoT
Resilient IoT data exchange using SDN and edge resources; 1st place at the Linux Foundation Akraino 5G MEC Hackathon and committer on an LF Edge blueprint.