Exploratory AI prototype for skin-analysis user experience, visual input handling, and safety-first product design.
Derm-Ai V1 is a product prototype exploring how AI-assisted visual analysis can be presented to users in a responsible, understandable, and controlled way.
This project is not positioned as a medical diagnosis tool. The focus is product thinking: how sensitive AI outputs should be framed, how user trust should be protected, and how a mobile experience can guide users without overstating certainty.
| Area | What Derm-Ai V1 Explores |
|---|---|
| Visual input | Capturing and preparing user-provided images for AI-supported analysis |
| User guidance | Helping users understand what kind of photo/input is suitable |
| AI interpretation | Presenting AI-generated observations with careful language |
| Safety boundaries | Avoiding diagnostic certainty and keeping outputs informational |
| Mobile UX | Designing a focused Android experience for sensitive AI interactions |
| Trust design | Making uncertainty, limitations, and next steps visible |
AI products in sensitive domains need more than model capability.
They need careful product boundaries, understandable outputs, and user experiences that do not create false confidence. In health-related contexts especially, speed and automation are not enough. The product must communicate uncertainty clearly and avoid replacing professional judgment.
Derm-Ai V1 explores that product challenge through a mobile prototype.
flowchart LR
A[User image input] --> B[Input quality guidance]
B --> C[AI-supported visual analysis]
C --> D[Structured observations]
D --> E[Safety-aware explanation]
E --> F[User-facing next-step guidance]
| Product Risk | Design Response |
|---|---|
| User treats output as diagnosis | Use informational language and avoid diagnostic certainty |
| Poor image quality affects result | Add input guidance and suitability checks |
| AI output appears too authoritative | Show observations, confidence boundaries, and limitations |
| Sensitive user context is misunderstood | Keep human/professional judgment central |
| Over-automation weakens trust | Design AI as a support layer, not a final authority |
| Capability | Product Value |
|---|---|
| Mobile-first experience | Tests how sensitive AI interactions work on Android |
| Image suitability checks | Reduces friction and improves input quality |
| Structured analysis output | Converts AI response into more readable user-facing signals |
| Safety-oriented copy | Keeps the experience cautious and responsible |
| Persistent analysis state | Supports continuity in user review flows |
| Brand and UI system work | Explores visual clarity for a health-adjacent AI product |
This project reflects my interest in AI products where user experience, trust, and safety are as important as technical capability.
Key product questions behind Derm-Ai V1:
| Product Question | Design Direction |
|---|---|
| How should AI outputs be shown in sensitive domains? | Use cautious, structured, and non-diagnostic language |
| How can users understand input quality requirements? | Guide image capture before analysis |
| How can uncertainty be communicated clearly? | Make limitations part of the user experience |
| How can AI support without replacing experts? | Position outputs as observations, not final decisions |
| How can mobile UX reduce confusion? | Keep flows focused, visual, and action-oriented |
flowchart TB
APP[Android App] --> UX[Guided Image Capture]
UX --> AI[AI Visual Analysis Layer]
AI --> OUT[Structured Observations]
OUT --> SAFE[Safety-Aware UX Copy]
APP --> STATE[Local / App State]
APP --> CONFIG[Configuration and Environment Setup]
| Layer | Stack |
|---|---|
| Platform | Android |
| Language | Kotlin |
| Build | Gradle / Kotlin DSL |
| Product Layer | Mobile UX flows, image handling, structured analysis views |
| AI Layer | Visual analysis workflow integration |
| Safety Layer | Input suitability checks and cautious output framing |
Exploratory public prototype and portfolio project.
Derm-Ai V1 is useful as a showcase for AI product thinking in sensitive contexts: trust, safety boundaries, mobile UX, and human-centered presentation of model outputs.
Derm-Ai V1 is part of my broader product focus around:
- AI-supported product experiences,
- responsible AI UX,
- human-in-the-loop product design,
- decision-support systems,
- turning model outputs into understandable user experiences.