The Challenge
Family caregivers are one of the most stressed, underserved populations in healthcare. They’re managing complex medical situations with minimal training, often while holding down jobs and raising kids. Most don’t have time for traditional therapy appointments.
My Role
Technical co-founder. Led a team of 10 engineers, 3 designers, and 5 care providers. Responsible for product architecture, clinical AI development, and customer discovery.
The Approach
We built a hybrid system:
- Empathetic virtual therapy bot delivering Problem-Solving Therapy (PST) and Motivational Interviewing—evidence-based approaches adapted for async text
- Smart response system for clinical team members—AI-drafted responses they could edit and send, dramatically speeding up care coordination
- Human-in-the-loop annotation ensuring training data quality from actual clinical experts
Empathic AI system capable of crafting nuanced responses to open-ended messages by inferring the user’s mental state—a state-of-the-art approach to emotional understanding in text.
Technical innovation: knowledge graph-based conversational agent for conducting daily check-ins and delivering structured therapy protocols.
Outcomes
- $49,600 CoMotion Innovation Gap Fund award
- Selected for NSF I-CORPS program (100+ customer discovery interviews)