Vocal Biomarker Analysis
I worked at Sonde Health, which is a biotechnology company that uses vocal biomarkers to unlock speech as a vital sign. Sonde’s technology allows people to speak into their phone and receive a real-time analysis of their health, such as respiratory conditions. While I was at Sonde, I did the following work:
Built machine learning & AI models using Python to detect vocal biomarkers consistent with respiratory illness, including COVID-19, asthma, and COPD from 6-seconds of speech.
Synthesized insights from NLP, time-series analysis, and linear & non-linear modeling to analyze physiological data, user demographics, and health information
Employed Design Thinking principles to guide the creation of Sonde’s mental health product, including developing surveys and conducting interviews with users
Led the Research-as-a-Service division by conceptualizing new data science projects, writing white papers and peer-reviewed articles, and mentoring data analysts and research interns
Two of the projects I worked on included white paper deliverables:
Nature wrote an article summarizing how vocal biomarkers can detect Covid-19