CuraSense-AI

Inspiration

💡 Inspiration

Healthcare access in remote regions like Nepal is often hindered by three major barriers: language gaps, lack of internet connectivity, and data privacy concerns. I was inspired to build CuraSense-AI to empower users in these areas with a private, multilingual "Health Agent" that works entirely offline, ensuring no one is left behind regardless of their location or language.

⚙️ How it works

CuraSense-AI acts as a local Triage Agent. It uses sophisticated on-device logic to analyze symptoms and provide a percentage-based probability for over 100 diseases.

• Interoperability: Designed with the MCP (Model Context Protocol) mindset, the app's diagnostic engine is built to eventually share structured data via FHIR standards, allowing it to "talk" to clinician systems.
• Risk Assessment: A dynamic "Risk Meter" categorizes health urgency into Low, Moderate, or High levels.
• Security: All processing is done locally to ensure total data sovereignty—no data ever leaves the device.

🛠️ How I built it

I developed this application using a modern Android stack:

• Kotlin & Jetpack Compose: For a fast, responsive, and multilingual UI.
• Room Database: To handle offline data persistence and secure medical record logs.
• Prompt Engineering: Optimized LLM logic for accurate symptom-to-disease mapping.

🚧 Challenges I faced

The biggest challenge was achieving high-accuracy diagnostics in an offline environment. Optimizing the symptom-matching logic to run without a server handshake required extensive testing and refined data structuring.

🧠 Accomplishments that I'm proud of

I am proud of creating a tool that can switch instantly between English, Nepali, Hindi, and Spanish, making complex health information accessible to non-English speakers.

📖 What I learned

I deepened my understanding of healthcare data standards and the importance of Edge AI (On-device AI) in building resilient, private healthcare solutions.

🚀 What's next for CuraSense-AI

My next goal is to fully integrate the Prompt Opinion Marketplace tools to allow this agent to seamlessly transfer secure data to professional FHIR servers for doctor consultations.

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for CuraSense-AI

Built With

• android
• fhir-standards
• jetpack-compose
• kotlin
• material
• mcp-(model-context-protocol)
• on-device-ai
• room-database