Clyvera — AI Health Assistant
A conversational health assistant that helps users understand symptoms, track conditions and organise medical records. I built the AI layer: retrieval over the medical knowledge base, the LLM integration and prompting, and the on-device data architecture that keeps health records off third-party servers.
Key Features
Retrieval over a knowledge base of 1,600+ health topics across 200+ medical categories
Conversational symptom analysis and personalised health insights
On-device data architecture built to meet HIPAA, GDPR and FDA requirements
Medication and appointment reminders
Caregiver tools for family health monitoring
The Challenge
Health AI removes the easy option. You cannot simply forward a user's symptoms to a third-party API, because the moment you do, protected health information leaves your control and HIPAA and GDPR obligations follow it. The architecture keeps records on the device rather than server-side, and the retrieval layer works against a curated knowledge base of 1,600+ health topics so answers are grounded in reviewed material instead of model recall.