How a Master’s student at Uppsala University transformed a research idea into a family-aware AI coaching system for people with Type 2 diabetes and their families — combining personalized support with privacy by design.
Meet the Health Builder
Leon Saidavi – Master’s Student, Uppsala University
Leon developed FamilyCoach Connect, an AI Diabetes Coach designed to support both people living with Type 2 diabetes and the family members who help care for them.
Using Arkus, he transformed a research concept into a functional AI solution, demonstrating how Health Builders can turn complex health AI ideas into working solutions faster.
The problem
More than 500 million people live with Type 2 Diabetes globally. But managing diabetes rarely affects only one person.
Family members often help with routines, lifestyle changes, motivation, and everyday decision-making — yet there are very few structured tools designed to support both patients and the people caring for them between clinical appointments.
That challenge became the starting point for FamilyCoach Connect:
a family-aware AI coaching system designed to support both patients and caregivers through personalized, context-aware interactions.
But the project quickly revealed another challenge: building advanced AI systems for health is much harder than it sounds.
The real barrier wasn’t the idea
The vision was clear. The technical reality was not.
To make the experience meaningful, the system needed:
- separate AI flows for different users
- session continuity
- personalized health context
- privacy-aware architecture
- structured workflows
- scalable agent logic
Building that infrastructure from scratch would normally require months of backend engineering work.
For a solo researcher, that complexity can become a major bottleneck. And this is exactly the kind of barrier many Health Builders face today: the ideas already exist — but turning them into real, working solutions remains difficult.
Building with Arkus
Using Arkus, FamilyCoach Connect moved from concept to working prototype dramatically faster than traditional development approaches would allow.
Instead of spending months building AI infrastructure, the focus could remain on:
- research
- coaching design
- user experience
- health outcomes
Arkus handled the heavy AI infrastructure layer, making it possible to quickly build, test, and iterate on the system.
The result was a multi-role AI experience with:
- separate patient and supporter coaching flows
- personalized interactions
- structured workflow logic
- session continuity
- privacy-by-design architecture
What was built
1. Dual-role AI coaching
FamilyCoach Connect includes two separate AI experiences:
- one for the patient
- one for the family supporter
Each role has different contexts, permissions, and coaching goals — powered by distinct Arkus agent flows.
2. Privacy by design
The patient remains in control of what information is shared.
Supporter agents are structurally restricted from accessing sensitive health information, helping maintain privacy boundaries throughout the experience.
3. Personalized health context
Health information, such as diabetes type, treatment, lifestyle, goals, and communication preferences, can be injected into conversations, making every interaction more relevant and personalized.
Why Arkus made a difference
Several aspects of Arkus were critical to the project.
Fast prototyping
What would traditionally take months of backend engineering work took weeks.
Arkus significantly reduced the gap between idea and working system.
Faster iteration
Testing and refining AI behavior became much easier.
Different approaches could be explored quickly without rebuilding infrastructure from scratch.
From research to running application
Arkus handled:
- AI orchestration
- session continuity
- context management
- multi-flow routing
This allowed the project to move beyond a research concept into a functional application.
Why this matters
FamilyCoach Connect demonstrates something increasingly important in health AI: meaningful health solutions should not require massive engineering teams to exist.
Researchers, clinicians, founders, patients, and innovators already understand many of the problems worth solving.
What’s often missing is accessible infrastructure that allows those ideas to become a reality.
This is the vision behind Arkus: empowering more people to become Health Builders.
What comes next
The project is evolving into Aura — a broader AI-powered health companion experience built on the same Arkus foundation.
The long-term vision includes:
- continuous diabetes support
- personalized health insights
- wearable integrations
- AI-assisted pattern recognition
- doctor-ready summaries
The goal is to create a more supportive and understandable health experience between clinical appointments.
From idea to impact
FamilyCoach Connect is more than a prototype. It’s an example of what becomes possible when builders can focus less on infrastructure — and more on solving meaningful health challenges.
With Arkus, the idea became a working solution.