Explored the design of context-aware assistive technologies for meal preparation using a user-centered iterative design process. Through three iterative phases of design and feedback, evolving from low-tech lightbox to a digital screen, we gained insights into managing diverse contexts and personalizing assistance through collaboration with older adults with MCI and their care partners.

Explored users' successful and unsatisfactory experiences with assistance from an LLM-based conversational agent, Mango Mango, during cooking.

Explored the impact of game fidelity on player empathy using the commercial reminiscing game Brukel, comparing a high-fidelity 3D version with a low-fidelity text-based version.

Explored user interactions with LLM-based voice assistants during a complex task -- cooking. We identified key verbal & nonverbal behaviors, and developed an analytical framework in three dimensions: 1. Key behavior characteristics (verbal & nonverbal), 2. Interaction stages (exploration, conflict, integration), 3. Stage transitions during the task.

Introduced an application that enables users to create their own playing cards for use in evocative storytelling games that generate unique card titles, cards suits, imagery, and poetry based on the user’s input to preserve their digital cards in an online repository and print them for tabletop game play use.

Explored whether a game design with higher fidelity and a specific type of scene would correlate with a better player experience compare to low-fidelity game using a commercial gaminiscing game, Brukel.

Catch The Butterfly is a narrative game exploring the real, lived experiences of an immigrant. This game aims to provide insights to game designers interested in the gaminiscing method, and to promote empathy and understanding of immigrants.

Explored the utilization of the gaminiscing method in the design of a narrative-driven game, where every game mechanic is derived from the authentic stories of an immigrant.

Introduced a personalized calendar organization tool that enhances student scheduling by learning from human preferences, implemented using and comparing two reinforcement learning algorithms: Linear Upper Confidence Bound and Preference Perceptron.

Explore the need for explainable AI (XAI) techniques to enhance the interpretability of black-box models. This project aims to reimplement the Local Interpretable Model-agnostic Explanations (LIME) method on a religion dataset to investigate its effectiveness compared to a baseline random K-features explanation similar to the original study.