CHARM - AI-Native Community Health Toolkit

[Health/AI]

Tech Stack

Android, Jetpack Compose, MVI, MVVM, Kotlin, React, JS/TS

Organization(s)

Causal Foundry

m2m (mothers2mothers)

Resources

CHARM - AI-Native Community Health Toolkit

Project Details

The Challenge

mothers2mothers (m2m) employs local women living with HIV/TB/... as Peer Mentors community health workers delivering frontline healthcare to over 16 million people across sub-Saharan Africa. Managing client information, appointments, and care delivery across low-connectivity, high-pressure environments was creating real gaps in care quality and Peer Mentor effectiveness.

The Solution

I led a cross-functional team from requirements gathering to MVP launch, building CHARM an AI-native platform that centralizes clinical and behavioral data, empowering Peer Mentors to deliver personalized, informed care more efficiently. Built with an offline-first mindset and on-device LLM support, CHARM works reliably even in low-connectivity settings where internet access is inconsistent.

What I Built

Client Management: Streamlined workflows for managing client profiles, assessments, and care histories with intuitive, tap-first UX designed for non-technical users in the field

Appointment Management: Simplified calendar and scheduling system ensuring no client appointment is missed or overlooked

Performance Tracking: Real-time insights and metrics that help supervisors and Peer Mentors identify improvement areas and track outcomes

AI-Driven Insights: On-device intelligence surfaces the most relevant client information at the right moment, reducing cognitive load and supporting better decisions

Personalized Notifications: Timely, context-aware nudges keeping Peer Mentors on top of critical tasks and client needs

Impact

Deployed across m2m's network in sub-Saharan Africa supporting Peer Mentors serving communities with some of the highest unmet healthcare needs in the world.