💡 Inspiration
Skin problems are often ignored until they become serious.
In many regions, access to dermatologists is limited, expensive, or delayed.
We asked a simple but powerful question:
What if we could predict skin conditions before they become visible problems?
That idea became NeuroDerm AI.
🚀 What We Built
NeuroDerm AI is an AI-powered skin health platform that goes beyond detection.
- 📸 Users upload a skin image
- 🧠 AI analyzes multiple skin conditions (acne, pigmentation, dryness, etc.)
- 🔮 A Digital Twin model simulates how the skin evolves over time
- 📊 Users receive predictions, risk insights, and personalized recommendations We don’t just analyze skin — we predict its future. ## 🧠 How We Built It We designed NeuroDerm AI as a full-stack intelligent system:
- Frontend: Next.js + TypeScript + Tailwind (modern responsive UI)
- Backend: FastAPI with real-time inference pipeline
- AI Model: Fine-tuned DINOv2 Vision Transformer for multi-label classification
- Database: PostgreSQL for user + report tracking
- Storage: Cloud-based image handling (S3/Cloudinary)
- Authentication: Secure JWT-based system We also implemented a digital twin simulation layer to model different skin scenarios:
- Current habits
- Worsening conditions
- Optimized skincare ## ⚙️ Challenges We Faced This project pushed us technically and conceptually:
- 🧩 Integrating ML with real-time frontend experience
- ⚡ Reducing inference time while maintaining accuracy
- 🔗 Fixing API communication & CORS issues
- 🧠 Designing a meaningful “prediction” system instead of just classification
- 📦 Managing full-stack deployment and environment configurations Every challenge forced us to rethink and improve the system architecture. ## 📚 What We Learned
- Building AI is not just about models — it’s about experience and usability
- Real-world impact requires speed, clarity, and trust
- Full-stack integration is harder than individual components
- The difference between a project and a product is execution quality ## 🌍 Impact NeuroDerm AI has the potential to:
- Improve early detection of skin conditions
- Reduce dependency on immediate specialist access
- Enable continuous skin monitoring
- Empower users with actionable health insights ## 🔮 Future Scope
- 📱 Mobile app integration
- 🧬 Personalized skincare recommendations using lifestyle + history
- 🏥 Integration with dermatologists for clinical validation
- 🌐 Deployment for rural and low-access healthcare systems
## 🏁 Conclusion
NeuroDerm AI is not just a diagnostic tool.
It is a step toward predictive, personalized, and accessible healthcare.
We built this in a hackathon —
but we believe it can scale far beyond it.
Built With
- cloudinary
- dinov2
- docker
- fastapi
- jwt-authentication
- next.js
- postgresql
- python
- pytorch
- react
- redis
- tailwind-css
- typescript



Log in or sign up for Devpost to join the conversation.