About Me

I am passionate about leveraging innovative AI technologies for positive societal impact, promoting equitable opportunities, and driving meaningful change toward societal well-being. My main interest is about leveraging non-verbal cues to gain deeper insights into human behavior patterns and understanding how humans think and act 👀🧠. With over 7 years of experience in machine learning, I specialize in projects involving eye gaze movement, behavioral data, time series data, and multi-modal approaches. I focus on integrating domain knowledge with AI to develop solutions that are both innovative and impactful, ensuring they address real-world challenges effectively. My work emphasizes optimization and applicability, particularly in Human-Computer Interaction (HCI) and multimodal AI systems. By leveraging behavioral insights, I enhance interaction design, adaptive interfaces, and user-centric applications, bridging the gap between foundational AI research and practical implementation.

Education

Ph.D. in Computer Science and Engineering

University of California, Santa Cruz (2021 - )

Master's in Computer Science

Kookmin University (Graduated in 2021)

Bachelor's in Computer Science

Kookmin University (Graduated in 2017)

Experiences

Research Scientist Intern

2024.6 - 2025.2

Meta, Redmond, WA

Reality Labs - Audio team

Applied Scientist Intern

2022.6 - 2022.9

Amazon, Seattle, WA

Computer Vision and NLP team
Early-fusion multi-modal model for product type classification

Publications

  • Reading with Screen Magnification; Eye Movement Analysis Using Compensated Gaze Tracks
  • Seongsil Heo, Roberto Manduchi, and Suzana Chung, In ACM Symposium on Eye Tracking Research & Applications(ETRA), MULTIPLEYE, April 2024
  • Eye Movement Analysis for Low Vision Readers Using a Full Screen Magnifier
  • Roberto Manduchi, Seongsil Heo, Suzana Chung, February 2024
  • Sage; A Multimodal Knowledge Graph-based Conversational Agent for Complex Task Guidance
  • Kaizhi Z., Jeshwanth B., Bhrigu G., Seongsil Heo, Vignesh, Dhananjay S., Winson C., Shree V., Wang E., 2023
  • Learning from Time Series
  • Seongsil Heo, In Kookmin University, Seoul, Korea, August 2021
  • Unsupervised Representation Learning for ECG-based Stress Detection (Best Paper Award)
  • Seongsil Heo, Jaekoo Lee, In IEMEK Symposium on Embedded Technology, May 2021
  • Stress Detection with Single Sensor PPG by Orchestrating Multiple Denoising and Peak Detection Method
  • Seongsil Heo, Sunyoung Kwon, and Jaekoo Lee, In IEEE Access, February 2021
  • PPG signal processing and comparison study with learning-based model for stress detection (Best Paper Award)
  • Seongsil Heo, Inkyung Kim, Sunyoung Kwon, Hyejin Lee and Jaekoo Lee, In Proceedings of Symposium of the Korean Institute of communications and Information Sciences, August 2020
  • A variation on Loss function of Deep Neural Networks for Facial Recognition
  • Seongsil Heo, Daehee Kim, Jaebin Lee and Jaekoo Lee, In Proceedings of Symposium of the Korean Institute of communications and Information Sciences, February 2020
  • Real-time Face De-identification in Visual Media using a Deep Neural Network for Object Detection
  • Seongsil Heo, Daehee Kim, Yonguk Kim and Jaekoo Lee, In Proceedings of Symposium of the Korean Institute of communications and Information Sciences, June 2019