Project Synapse

Inspiration

This project started from a very personal place. A close family member of mine had to see multiple doctors for a serious health issue, and we were trapped in a frustrating cycle. Each specialist gave their own advice—sometimes contradicting the others. We had to repeat the same medical history over and over, carry binders of test results from one appointment to the next, and ended up feeling more confused than reassured.

Watching them struggle through that process made me realise how overwhelming it is when the system itself adds to a patient’s burden. That very frustration sparked a simple question: What if technology could provide the clarity and efficiency that the system so clearly lacks?

That’s what inspired this project. We are utilising agentic AI to develop a tool that streamlines the logistical chaos of a hospital visit. By automating information transfer and consolidating medical data into a single, coherent view, we can empower doctors to diagnose more quickly and provide patients with what they were looking for all along: a clear, confident path to recovery.

What it does

So what we did is build an Agentic AI-powered personal health coordinator. At its core is a master agent that brings together a network of specialised AI sub-agents—such as cardiology, neurology, dermatology, and endocrinology. Each sub-agent interprets relevant patient data, follows medical guidelines, and generates its own recommendations.

The master agent consolidates all sub-agent recommendations into a single, coherent report that flags contradictions, highlights risks, and organises information into three tiers: urgent needs, long-term monitoring, and lifestyle changes. This system provides clarity for patients and serves as a longitudinal health record for doctors, eliminating redundant paperwork. To ensure secure and intelligent processing, the platform uses an MCP server for communication, FHIR for structured data, and a consensus engine to weigh the alignment of each medical opinion. This integrated approach provides patients with consistent second opinions and gives doctors a streamlined view of all prior inputs, leading to faster decisions and greater trust in the care plan.

Built With

• apis
• cloud-services
• databases
• fhir
• frameworks
• google-cloud
• huggingface
• langchain
• numpy
• openai
• pandas
• platforms
• postgresql
• pubmedqa
• python
• pytorch
• react
• scikit
• sql
• streamlit
• supabase