Outcome-based weighting and validation of a wearable-derived functional aging biomarker - BC-1021
Project type: ResearchDesired discipline(s): Engineering - biomedical, Engineering, Kinesiology, Life Sciences, Computer science, Mathematical Sciences
Company: UnDun Longevity Innovations
Project Length: 4 to 6 months
Preferred start date: As soon as possible.
Language requirement: English
Location(s): Burnaby/Vancouver, BC, Canada
No. of positions: 2
Desired education level: CollegeUndergraduate/BachelorMaster'sPhDPostdoctoral fellowRecent graduate
Open to applicants registered at an institution outside of Canada: No
About the company:
UnDun Longevity Innovations is an early stage health technology company building a continuous functional aging monitoring system. The company has developed UHLI (UnDun Health Longevity Intelligence), a device-agnostic computational framework that derives Intrinsic Capacity domain scores from continuous consumer wearable data (Apple Health, Oura, Whoop, Fitbit, Garmin, Ultrahuman, and Android Health Connect) using individual-level (n=1) baseline modeling rather than population norms. UnDun is recognized by NVIDIA Inception, AWS, and Google Cloud, and maintains active research partnerships with leading longevity scientists, including LiveBeyond (co-founded by Brian Kennedy, former CEO of the Buck Institute for Research on Aging) and Professor Andrea Meier of Singapore. The company currently has over 30 active users generating longitudinal wearable data and is focused on building a clinical reporting layer for longevity clinics, alongside ongoing scientific validation of its scoring framework.
Describe the project.:
UnDun Longevity Innovations has developed UHLI (UnDun Health Longevity Intelligence), a device-agnostic computational framework that derives Intrinsic Capacity domain scores from continuous consumer wearable data using individual-level (n=1) baseline modeling. The framework currently maps twelve physiological indicator variables onto four functional domains (locomotion, vitality, cognition, and psychological health) using equal weighting within each domain.
The goal of this research project is to develop and validate outcome-based weighting approaches for the UHLI framework using machine learning and statistical optimization techniques. The research will explore whether differential weighting of physiological indicators, informed by clinically meaningful outcomes such as frailty, disability risk, and functional decline, improves the predictive validity and responsiveness of the model compared to the current equal-weighting approach.
Additional research questions include: (1) how wearable-derived Intrinsic Capacity trajectories relate to established biological aging biomarkers, including DNA methylation-based aging clocks; (2) how to detect and correct for skewed individual baselines arising from atypical calibration periods; and (3) how to improve signal-to-noise discrimination in distinguishing meaningful physiological change from normal day-to-day variation, including illness, travel, and other confounding factors.
The intern(s) will work with UnDun's existing longitudinal dataset (30+ active users with up to 4 months of continuous wearable data) to conduct exploratory data analysis, develop and test weighting and optimization models, and produce recommendations for the next iteration of the UHLI scoring architecture.
Required expertise/skills:
Strong background in biostatistics, machine learning, or computational biology. Proficiency in Python (pandas, scikit-learn, numpy) or R for statistical modeling and data analysis. Experience working with longitudinal or time-series health data. Familiarity with wearable device data (HRV, sleep, activity metrics) is an asset. Familiarity with aging biomarkers, Intrinsic Capacity, or epidemiological methods is an asset. Strong written communication skills for documenting methodology and findings.
