Current direction / Healthcare
AI for Health Signals & Care
I work on models and methods for health signals, from cardiopulmonary audio to ECG and other time series. I also collaborate with hospitals to connect research questions with clinical data and practice.
The research thread
Health signals are rich, noisy, and connected to decisions that matter. I am interested in methods that can learn from them, reason across modalities and tasks, and stay grounded in the evidence available. Working with hospitals helps shape the questions, data, and evaluation around real needs.
Projects and papers
An audio language model for cardiopulmonary analysis across a range of clinical tasks.
A scientific simulator authoring harness evaluated on cardiac and respiratory audio, PPG, and ECG. The full research story lives under Agentic Harnesses for AI4S.
Ongoing collaboration
I am working with hospitals on health data and research questions. I welcome conversations with healthcare professionals about problems where audio, ECG, or other time series might support new methods and careful evaluation.
If you work with clinical signals or hospital data, see Connect for ways to get in touch.