I design conversational agents and digital health tools that help people living with serious mental illness manage their care, and that bring the people who support them into it.

I am a Postdoctoral Fellow at the Center for Technology and Behavioral Health at Dartmouth, working where health, technology, and artificial intelligence meet. I am currently building next-generation AI-driven mental health interventions: adaptive, personalized support systems for undergraduate well-being that combine generative AI with passive sensing and multimodal data analysis.

I earned my Ph.D. in Informatics from the College of Information Sciences and Technology at Penn State, where I worked with Dr. Saeed Abdullah in the Well-being and Health Innovation Lab. My doctoral work focused on designing, developing, and deploying conversational agents that deliver therapeutic content and support self-management for PTSD. I also built deep learning models that assess symptoms such as PTSD and dementia from text and speech, developed privacy-preserving financial technologies built on open banking data to reduce the risk of financial harm for people with mental illness, and studied how care partners support financial management for people with bipolar disorder.

Before Penn State, I received a B.S. in Computer Science and Engineering from Ewha Womans University and an M.S. in Computer Science from the University of California, Irvine, where my thesis predicted everyday stress levels from wearable sensors.

  • Digital mental health
  • Conversational agents
  • Human–computer interaction
  • Collaborative care
  • Passive sensing
  • Health FinTech

News

  1. Presented “Large Language Models for Context-Aware Conversational Decision-Making in Digital Mental Health” at the Technology in Psychiatry Summit (TIPS) 2025, in the symposium on building intelligent mental health apps.

  2. Joined the Center for Technology and Behavioral Health at Dartmouth as a Postdoctoral Fellow.

  3. Officially earned my Ph.D. in Informatics.

  4. Successfully defended my dissertation.

  5. Our paper on collaborative FinTech for the financial lives of people with bipolar disorder was accepted to the CHI 2025 Workshop on the Future of Money and HCI.

  6. Selected to attend the Grace Hopper Celebration of Women in Computing.

  7. Our paper on the acceptance and feasibility of a conversational agent for PTSD self-management was accepted to SAGE Digital Health.

  8. Our paper evaluating a conversational agent for PTSD self-management with clinical experts was accepted to JMIR Formative Research.

  9. Passed my comprehensive exam.

  10. Presented our work on diagnosing dementia from text and speech data at the ICDS Symposium.

  11. Our paper on classifying dementia from text and speech data was accepted to Multimodal AI in Healthcare.

  12. Won the W3PHIAI-22 Data Hackallenge with a model that classifies dementia from text and speech data, and presented it at the 2022 AAAI Workshop on Health Intelligence.

  13. Served as a student volunteer at UbiComp/ISWC 2021.

  14. Our paper on designing a conversational agent for people with PTSD was accepted to the UbiComp/ISWC Adjunct Proceedings.

  15. Attended the 2021 CRA-WP Grad Cohort for Women.

  16. Started my Ph.D. at the College of Information Sciences and Technology at Penn State, advised by Dr. Saeed Abdullah.

Publications 11 papers, 2 Best Paper Awards

2025
2024
2023
2021
2020
2016

Research experience

  1. Aug 2025 –
    present

    Center for Technology and Behavioral Health

    Department of Biomedical Data Science, Geisel School of Medicine, Dartmouth

    Postdoctoral Fellow

    • Designing and evaluating adaptive, personalized AI-driven support for undergraduate well-being, using generative AI, passive sensing, and multimodal data.
  2. Aug 2020 –
    Aug 2025

    Wellbeing & Health Innovation Lab

    College of Information Sciences and Technology, Penn State

    Graduate Research Assistant

    • FinTech for bipolar disorder. Built privacy-preserving tools on open banking data to flag risky financial behavior and offer personalized interventions, and studied how care partners support financial management.
    • PTSDialogue. Developed a conversational agent that delivers therapeutic content and supportive accountability for PTSD, and assessed its feasibility and acceptability.
    • ACOUSTIC. Built a deep learning model to classify dementia from text and speech data.
  3. Jun 2018 –
    Jul 2019

    Dutt Research Group

    Department of Computer Science, UC Irvine

    Graduate Research Assistant

    • Proposed a stress monitoring system that predicts everyday stress levels from PPG, ECG, and GSR sensors on a wearable device.
  4. Aug 2015 –
    Jun 2016

    Computer Architecture and System Design Laboratory

    Department of Computer Science and Engineering, Ewha Womans University

    Undergraduate Research Assistant

    • Built a multimodal pedestrian counting system with two algorithms (Adaptive-DBSCAN and a pedestrian algorithm) and two new hardware devices (F-pad and Sonic-Bar).

Projects

  • FinTech to support people with bipolar disorder

    May 2024 – ongoing

    User-centric, privacy-preserving financial tools that use open banking data to spot problematic financial behavior and deliver personalized interventions.

  • Detecting mental health symptoms from text, enhanced by ChatGPT

    Jan 2023 – ongoing

    A deep learning model that assesses mental health symptoms from text with 90.4% accuracy, to be strengthened with ChatGPT-generated labels once their accuracy is evaluated.

  • PTSDialogue

    Aug 2020 – 2024

    A conversational agent that delivers therapeutic content and supportive accountability to sustain engagement for people with PTSD.

  • Automatic classification of dementia

    Feb – Mar 2022

    An ensemble of two deep learning architectures for text and speech that classified dementia versus healthy controls with 89.8% accuracy. Won the W3PHIAI-22 Data Hackallenge.

  • WiSDoM: guiding internet users toward safer privacy decisions

    Jan – Jun 2021, Demo paper (PDF)

    A Chrome extension, powered by a BERT-based disclosure detection model, that shows users what they are disclosing online.

  • Ubiquitous stress monitoring with wearables

    Jun 2018 – Jul 2019

    A healthcare IoT system that predicts stress levels in everyday settings from PPG, ECG, and GSR signals on Shimmer devices.

Education

  1. 2020 – 2025

    Ph.D. in Informatics

    Pennsylvania State University

  2. 2017 – 2019

    M.S. in Computer Science

    University of California, Irvine

    • Advisor: Nikil Dutt
    • Thesis: Objective Stress Monitoring Based on Wearable Sensors in Everyday Settings
  3. 2012 – 2017

    B.S. in Computer Science and Engineering

    Ewha Womans University

    • Advisor: Dong Sub Cho
    • ABEEK accredited (Accreditation Board for Engineering Education of Korea)