Eugene Bagdasarian
Scholar

Eugene Bagdasarian

Google Scholar ID: _MfoOC8AAAAJ
UMass Amherst, Google
ML SecurityML PrivacyContextual IntegrityTrustworthy AIAI Safety
Citations & Impact
All-time
Citations
5,048
 
H-index
15
 
i10-index
18
 
Publications
20
 
Co-authors
31
list available
Resume (English only)
Academic Achievements
  • Apple Scholars in AI/ML Fellowship
  • Digital Life Initiative Fellowship
  • Usenix Security Distinguished Paper Award
  • Schmidt Sciences AI Safety Grant
  • Published papers at top venues including CCS'24, HotOS'25, USENIX Security'20/24/25, AISTATS'20, S&P'22
  • Proposed frameworks such as AirGapAgent (CCS'24), Conseca (HotOS'25), Terrarium, Backdoors101, Mithridates
  • Contributions in multimodal security, LLM attacks, backdoor attacks, and reasoning attacks
Background
  • Assistant Professor of Computer Science at UMass Amherst CICS
  • Research focuses on security and privacy attack vectors for real-world AI systems
  • Co-leads the AI Security Lab at UMass with Amir Houmansadr
  • Co-leads the AI Safety Initiative with Shlomo Zilberstein
  • Senior research scientist (part-time) at Google working on agentic privacy and security