I am working as a PhD student with Professor Christian Holz at the Sensing, Interaction & Perception Lab at ETH Zurich, where I apply statistics and statistical machine learning to large medical datasets.
My research focuses on biomedical time series for disease modeling. I first explored perceived health through wearable sensor data in intensive longitudinal studies (see my publications on perceived health). Since then, I have developed new methods to extract information from wearables (e.g., Nightbeat) and used UK Biobank data to model disease and mortality risks at the population scale (currently under review). Now, I’m building interpretability methods for irregular time series models, with applications to wearables data and electronic health records.
Below are some of the most recent publications I’ve been involved in. You can check out a full list of my publications here.
There are a few common themes: interpretable modeling, mobile health, perceived health, and human computer interaction.
Most of my past projects involved interpretable modeling techniques to better understand the outcome of interest rather than simply predicting it. A few publications focus on perceived health, such as fatigue or sleep quality, and I’ve been a sidekick on a few publications in human computer interaction, where I mainly contributed to the (interpretable) statistical analysis.