Ph.D. student
Aug 2024–present
Master’s · May 2023–Aug 2024
Master’s student
Jan 2025–present
Mia’s research on the reliability of LLM-based fire detection and data analysis was selected for NSF EPSCoR UROP funding. The project examines how indirect prompt injection through external information can distort wildfire risk assessments and response recommendations. UNR research feature
Jiahao received the Outstanding Graduate Student Researcher Scholarship from the University of Nevada, Reno.
Jiahao and Zikai received first-place recognition in the UNR Cybersecurity Conference poster competition. This marked the team’s third consecutive first-place finish, spanning 2023–2025.
Jiahao received MobiHoc 2025 student travel grant in October. Jiahao and Zikai each received NSF EPSCoR HDRFS support for student publications and conference travel in August.
Jiahao's paper AlignIns, on detecting backdoor attacks in federated learning, was selected as a Highlight at CVPR 2025. The method identifies suspicious model updates by examining their directional alignment. Jiahao also received CVPR 2025 travel support in April.
Jiahao received the 2024 Nevada WateReuse Research Grant as principal investigator. His project investigates robust water-leak detection using federated learning with sparsified local models. WateReuse award recipients
Zikai received an NSF-supported travel grant to attend the IEEE Conference on Communications and Network Security (CNS 2023).
I am recruiting motivated Ph.D. students with assistantship support. Our interests include AI security, data privacy, efficient distributed/federated learning, multimodal LLMs, agentic AI, and reinforcement learning.
Prospective interns are welcome to contact me with their research interests and relevant experience. Please indicate that you are seeking a remote internship.
Email me with your CV, transcripts, TOEFL and GRE scores, and a research statement. Indicate the opportunity you are seeking and how your interests connect with our work. Relevant papers, projects, or code are welcome.
Given the volume of inquiries, I may be unable to respond to every message. I will contact applicants whose interests and experience align with current opportunities.
Undergraduate and graduate students interested in LLM security, AI agents, or real-time scene analysis with vision-language models are welcome to email me or stop by my office. Students interested in AI security may also consider CS 791: CyberAI, offered in spring.
Student Team: John Althoff, Aiden Coss, and Karam Alkherej
Developed a web platform for evaluating large language models against prompt injection, jailbreaks, and unsafe outputs. The platform connects to models through APIs, runs automated adversarial tests, and generates reports to support repeatable security assessment.
Student Team: Christine Angela Barlaan, Evelynne (Evie) Nivera, and Matthew Osorio
Built a voice-driven detective interrogation game with AI-generated dialogue and adaptive story progression. Optional wearable-sensor feedback allows the virtual detectives to respond to the player’s physiological signals.
Student Team: Zachary Strazi, Kristian Konstantinov, Cody Long, and Jacob Ayers
Investigated adversarial audio perturbations that cause speech recognition systems to misinterpret spoken commands. The team evaluated models including Whisper and Wav2Vec2 and built a Raspberry Pi demonstration to study attacks on voice-controlled devices.