Yenho Chen

About Me

I am a Machine Learning Ph.D. Candidate at Georgia Tech, advised by Dr. Christopher Rozell. I’m interested in building foundation models for health and science. My work centers on pretraining strategies that build domain structure into learned representations, so that models generalize better and need fewer labels. In my thesis, I develop this for health time series by building dynamical systems structure into self-supervised pretraining, ranging from small interpretable models for low-data settings to large-scale pretraining on single- and multi-sensor wearable data.

Outside of my PhD, I have worked on a range of research problems through industry internships. At IBM Research, I developed a pretraining approach for designing new CAR-T therapy variants, and built an agentic evaluation harness to benchmark LLM agents on protein research tasks. At Blackstone Real Estate, I focused on multimodal forecasting models for hotel demand based on demand history (time-series) and market reports (text). At Travelers, I investigated domain reweighting strategies for continued pretraining of LLMs.

Prior to my PhD, I was a research fellow in the Machine Learning Team at the National Institute of Mental Health, where I built statistical tools for neural signal analysis.

I am currently on the market for research and applied scientist roles.

Research Interests

• Machine Learning: Foundation Models, Wearable AI, Self-Supervised Learning, Time-Series, Dynamical Systems Theory
• Applications: Health Time-series (EEG, ECG, PPG, HAR, Wearables), Large Language Models (LLMs), Protein Language Models

News

Publications

Reviewer Service

• ICML. 2025-2026
• ICLR. 2025-2026
• NeurIPS. 2024-2025
• Time-Series 4 Health Workshop @ NeurIPS. 2025
• Journal of Neural Engineering. 2023