Grin

Decoupled service communication through an event backbone with fan-out patterns. Services publish domain events that trigger downstream workflows without direct coupling, enabling independent scaling and deployment of each microservice.

State machine orchestrating the full patient treatment journey from initial assessment through active treatment to completion. Each transition triggers appropriate notifications, record updates, and provider alerts across the system.

Unified notification system delivering messages across SMS, email, push notifications, and in-app channels. Channel selection is driven by patient preferences and message urgency, with delivery tracking and retry logic for each channel.

Integration layer connecting 11 ML services for computer vision analysis and natural language processing. The pipeline handles image ingestion, model inference, confidence scoring, and result delivery back to the clinical workflow in real time.

WebSocket-based infrastructure enabling live updates between providers, patients, and the platform. GraphQL subscriptions push treatment changes, appointment updates, and diagnostic results to connected clients without polling.

Internal SDK consumed by all microservices, providing standardized patterns for authentication, event publishing, database access, and error handling. Reduces boilerplate and enforces consistency across the 28-service architecture.