Building a Multi-Agent Medical Triage System with Angular, Firebase & Genkit

A deep-dive into building a production-ready, multi-agent AI medical application using Angular, Firebase Cloud Functions, and Google Genki, powered by Gemini 2.5. A single Large Language Model (LLM) can answer questions, but real-world complexity in heavily regulated fields like healthcare requires specialized boundaries. You wouldn't ask a receptionist to diagnose an illness, and you wouldn't ask a doctor to figure out the exact GPS coordinates of a clinic. Similarly, we can create specialized AI agents that focus on one specific task and coordinate with each other to solve complex user flows safely. Overview For this project, I built AfyaSense , an AI-powered medical triage assistant. It helps users evaluate their physical symptoms, determines a medical risk level ( EMERGENCY , CLINIC , or SELF_CARE ), and dynamically locates nearby healthcare facilities through a single chat interface. What makes it interesting from an engineering perspective is the multi-agent routing architecture