AI today is good at predicting your next click — but not whether you are about to burn out. A digital twin that flags rising blood pressure but cannot link it to weeks of after-hours work, financial strain, or fading social ties has missed the point. Most digital health systems track a single signal in isolation and treat wellbeing as a fixed, narrowly-scoped concept. This project builds a Wholebeing Digital Twin — an AI model that fuses multimodal data from wearables, smartphones, and short self-reports across several interconnected life domains (physical, mental, social, occupational, environmental, financial, and digital) to deliver a context-aware, time-sensitive view of holistic wellbeing. Three intertwined research challenges drive the work: (i) scoring each domain from noisy, sparse, multimodal signals where some dimensions are easy to sense passively (physical, digital) while others are not (social, emotional); (ii) prioritizing domains dynamically, since their importance shifts with age, role, culture, and life events — a child does not weigh occupational satisfaction the same way an adult does; and (iii) modeling inter-domain links so the twin can explain why wellbeing is changing, not just that it is. The work is grounded on a pilot involving smartphone passive sensing, wearable devices, and ecological momentary assessment. The intern will help build the mobile data-collection app, develop a time-series pipeline that learns time-varying domain priorities, and contribute to a peer-reviewed publication.

1. A cross-platform mobile data-collection app integrating smartphone passive sensors, wearable device APIs (e.g., Fitbit), and ecological momentary assessment prompts. 2. A time-series modeling pipeline that ingests the multimodal streams and learns time-varying, per-domain wellbeing priorities, validated against baselines. 3. A final evaluation report and a co-authored research paper draft targeted at a top-tier IoT, pervasive computing, or digital-health venue.