PROJECTS

A human-in-the-loop AI system that adapts scent sequences to regulate sustained physiological arousal in high-stress driving scenarios. The project combines RLHF, physiological sensing, and interpretable scent–emotion representations to study non-intrusive, long-horizon affective intervention.

An AI-supported storytelling system that shifts intervention away from in-the-moment mediation toward pre- and post-session reflection. By visualizing emotional synchrony between parents and children, the project explores how generative AI can support parental agency, reflection, and emotional bonding.

An eyewear-based embodied AI system that translates visual scenes into spatialized audio narratives. NOEMA reframes perception from vision to sound, exploring how large language models and multisensory feedback can reshape attention, situational awareness, and everyday embodied experience.

A multimodal AI system that detects violent gameplay behavior through facial expression, audio, video, and interaction signals, triggering calibrated pain-based haptic feedback via a custom hardware device. The project investigates real-time behavioral regulation and ethical intervention through embodied feedback.

The Anemoia Device is a synthetic memory generator that uses generative AI to provoke nostalgia for a time you have never experienced. The work is an inquiry into memory malleability in an age of AI, proposing an alternative to conventional screen-based interaction through an intentional, embodied ritual that positions the user as an active co-author, rather than a passive consumer.

A generative AI creative support tool designed to align with how designers actually work: non-linearly, incrementally, and reflectively. The system integrates with existing 3D modeling workflows, capturing design progress over time and allowing designers to selectively request AI suggestions based on prior iterations, supporting exploration while preserving authorship and control.

A human-centered AI decision support system for sepsis care that examines how interface design shapes clinicians’ trust, interpretation, and use of model predictions. Through interface prototyping and user studies, the project investigates how AI can better support clinical reasoning in high-stakes, time-sensitive medical contexts.