Client
Role
Year
AI-powered rehabilitation platform
Selfit Medical is an AI-powered rehabilitation platform designed for hospitals, clinics, nursing homes, and eventually wellness centers and in-home care. Originally created for post-stroke patients, the system evolved - through research and user testing, into a broader solution for geriatric cognitive and motor rehabilitation.
I led the UX/UI design of the therapist interface and the on-floor exercise experience, shaping how therapists monitor patients, adjust sessions, and deliver personalized treatment plans supported by real-time motion analysis.
Transform rehabilitation into an intelligent, automated process
Rehabilitation is traditionally manual - Therapists observe movement, write notes, adjust exercises, and track progress themselves - all of which takes time, focus, and clinical effort.
Selfit’s mission was to transform this process into an intelligent, automated system that:
Recognizes patient performance
Detects difficulties therapists may miss
Suggests personalized treatment plans
Adapts exercises using AI and machine learning
Reduces therapist cognitive load
Works both on-site (clinics, hospitals) and eventually in the home or community setting
Rehabilitation Starts With Memory, Not Motion
At the start, Selfit’s vision centered on improving movement accuracy for post-stroke patients through real-time tracking. The first version delivered exactly that: a digital mobility-training experience with motion feedback and performance scoring. But once this version was deployed into real clinical environments, our testing surfaced deeper needs:
1. Cognitive decline was a major pain point
Mobility drills weren't the main bottleneck - Therapists also needed strong cognitive-motor combinations tools for memory, attention, sequencing, and problem-solving.
2. Therapists had no time for complex setup
The early version required manual configuration.
We needed something fast, reliable, and stress-free which led to the creation of Quick Mode - pre-built treatment templates that launch instantly.
3. Subtle patient issues often go unnoticed
Older adults especially display small signs of difficulty that are hard to track manually, so we had to make sure the system now identifies subtle issues therapists might miss, such as fatigue, symmetry, slower cognitive responses and reduced accuracy over time.
4. The system had to scale beyond stroke rehab
User interface had to work for older adults, meaning High contrast, large elements, minimal distractions - calm and supportive. The product needed to be extensible to support hospitals, clinics, nursing homes, wellness centers and in the future also home care.
This meant rethinking the platform from a mobility tool into a multi-domain therapeutic system.
I led UX/UI design across all phases, from concept to the pivoted second version
Version 1: Mobility-Focused
Exercise library & session builder
Therapist dashboard
Full on-floor visual system
Real-time movement tracking UI
Progress analytics
Version 2: Pivot to Cognitive + Geriatric Rehabilitation
Expanded task types (memory, sequencing, attention)
Updated patient models and task structures
Redesigned interface for older adults’ needs
AI insight visualization
New navigation & treatment flows
Quick Mode - instant session templates for time-pressed therapists
Combining the therapist app with the on-floor Exercise display
My design approach focused on the realities of clinical work - therapists juggling multiple patients and needing fast, clear, low-friction workflows. I streamlined flows, clarified patient overviews, and reduced cognitive load across the interface.
A key outcome was Quick Mode, enabling therapists to launch ready-made treatment sessions instantly, addressing the major time constraints revealed in user testing.
As the product shifted from mobility to cognitive-motor rehabilitation, I redesigned the floor-screen experience to be simple, high-contrast, and accessible for older adults. All of this had to be achieved under strict technical limitations, as the company was developing low-cost hardware to support the system.
Outcome
The product shifted from a mobility-only solution to a comprehensive cognitive-motor rehabilitation platform with a clearer value proposition and a broader clinical reach.
What I Learned
This project taught me how to transform complex therapeutic processes into calm, intuitive interfaces that clinicians can rely on.
User testing played a critical role, revealing gaps and opportunities no initial research could have predicted. Above all, I learned how thoughtful design can support better care by bringing together technology, empathy, and clinical insight.
