Inspiration
Healthcare’s fragmented data systems and delayed decision-making inspired HealthChain AI. We envisioned a future where AI cuts through medical complexity, blockchain ensures trust, and patients control their health narrative—no more siloed records or prescription errors.
What it does
Doctors: Query EHRs via AI (e.g., “Summarize Patient X’s allergies”) and receive instant, RAG-powered insights.
Pharmacies: Get AI-suggested drug alternatives, validated via blockchain approvals.
Patients: Track progress via an AI chatbot and securely share symptoms.
How we built it
AI/LLM: Fine-tuned Mistral-7B for medical RAG, using Perplexity’s APIs for multi-model routing.
Blockchain: Hyperledger Fabric for EHR access control + IPFS/S3 for encrypted data storage.
UI: React frontend with a UI-friendly interface to simplify blockchain interactions.
Challenges we ran into
RAG Model Selection: Tested 5+ LLMs (Llama2, GPT-3.5) before settling on Mistral-7B for accuracy/speed balance.
Solo Development: Juggled smart contracts, AI pipelines, and UI/UX design alone—time management was brutal.
Blockchain-AI Integration: Syncing real-time AI queries with on-chain permissions required custom REST APIs.
Accomplishments that we're proud of
Created an entire secured blockchain architecture in compliance with healthcare policies and compliances, also integrated different RAG models for reducing delays in the healthcare system.
What we learned
Blockchain isn’t just for crypto—it’s a privacy game-changer for healthcare. RAG models need domain-specific tuning (e.g., medical jargon handling).
Simplicity is key: hiding blockchain complexity behind AI/UX drives adoption.
What's next for HealthChain AI
AI Audit Trails: Log model decisions on-chain for transparency.
IoT Integration: Pull real-time data from wearables into EHRs.
Scale Globally: Partner with clinics to pilot in rural areas with limited healthcare access.
Log in or sign up for Devpost to join the conversation.