Healthcare information online is often confusing, overwhelming, or unreliable. I wanted to build something that helps people understand their symptoms clearly and safely, without replacing doctors. This inspired DoctorAI, an AI-powered symptom interpretation assistant that provides structured, easy-to-understand health guidance.
🧠 What I Learned
Working on DoctorAI taught me about:
Medical triage logic and red-flag symptoms
Prompt engineering for constrained, safe medical outputs
Using retrieval systems to ground LLM responses in facts
Designing clear UX for sensitive informationI built DoctorAI using:
An LLM workflow for symptom extraction and reasoning
A vector similarity system for mapping user symptoms to medical clusters
A backend API to structure outputs into JSON for consistent triage
A simple UI for users to input symptoms and view reports
Safety layers to ensure: no diagnoses, no prescriptions, and retrieval-grounded responses.
Log in or sign up for Devpost to join the conversation.