SELECTED WORK

SafePro: Evaluating the Safety of Professional-Level AI Agents
ICLR 2026AI Safety • Agents

Mirror: AI that knows you better than your mirror
CruzHacks WinnerComputer Vision • Mobile

Revoteen: helping students find
internships and opportunities
10k+ users70+ countries • 7+ media outlets

CCC: Enhancing Video Generation via MLLM Feedback
ICML 2025Video Generation • MLLMs

Smells Like Fire: Olfactory Cues in VR Evacuation Training
CHI 2026VR • HCI

Human-centered World Modeling for Multi-Agent AI
CHI 2025Multi-Agent • World Models
KIND WORDS

