Haojian Jin


I'm an Assistant Professor in Halıcıoğlu Data Science Institute (HDSI) and Department of Computer Science and Engineering (affiliate) at UC San Diego.

Human developers create risky computer systems that eventually affect human users. Our lab, Data Smith Lab, takes a human-centered approach to (1) help developers create systems with enhanced privacy and security features and (2) help users safeguard their privacy and security.

These two aspects often need to be codesigned so they do not interfere. For example, the incorporation of end-to-end encryption can make it hard for third parties to audit the privacy attributes of a smart camera.

I am recruiting PhD students and MS/undergrad research interns. See more details here.

Some Current Research Directions:
Effective Privacy Design

Effective Privacy Design: translate high-level privacy principles into specific low-level system design decisions.
Papers: [CHI'24] [CHI'22]
Software: [Work-in-progress];
Course: [Designing Privacy-Sensitive Data Systems]

Precision Privacy

Precision Privacy: understand individual users' privacy desires using data science methods and cognitive principles.
Papers: [CHI'24] [TOCHI'21]
Software: [Work-in-progress]

aimanifest

Superintelligence Safety: control AI models' behavior using human specified policies.
Papers: [CCS'24] [ICLR'25]
Software: Moderator, Game Arena

… and a bunch of earlier stuff:
modular privvacy flows

Modular Privacy Flows: implement fine-grained data access using operators and manifests.
Papers: [IEEE S&P 2022] [Dissertation summary] [IMWUT'18]
Software: Who knows what about us and why?, Peekaboo

software design cooking

Sensor-based Systems: Radio frequency, Audio, and Actuation.
Papers: [CACM'21] [MobiCom'19] [MobiSys'18] [IMWUT'17] [UIST'15] [UIST'15]
Patents: [11 Granted US Patents]