PAPPY-JOE/nurse-naija

NurseNaija ### Multilingual AI Triage Assistant for Community Clinics NurseNaija is a multilingual, AI-powered **medical triage assistant** designed to improve access to basic healthcare guidance in **community clinics and low-resource settings across Nigeria**. The system leverages **N-ATLaS (Nigeria’s national multilingual large language model)** to understand patient symptoms expressed in local languages and provide **safe, structured triage-level guidance** — without diagnosing or prescribing medication. ## 1. Problem Statement Many community clinics in Nigeria face: - Language barriers between patients and healthcare workers - Shortage of trained medical personnel - High patient volume with limited triage capacity As a result, **urgent cases are often delayed**, while non-urgent cases overwhelm limited resources. ## 2. Solution: NurseNaija NurseNaija acts as a **first-line triage assistant**, helping clinics and outreach programs: - Understand patient symptoms in **English, Hausa, Igbo, Yoruba, and Nigerian Pidgin** - Classify cases into **EMERGENCY, URGENT, or NON-URGENT** - Provide **immediate, non-diagnostic next steps** - Generate **structured triage outputs** that can integrate into clinic workflows The system is designed to be **safe, explainable, and deployable in constrained environments**. ## 3. Core AI Technology - **N-ATLaS (NCAIR1/N-ATLaS)** – multilingual reasoning & triage classification - **Nigerian-language ASR models** – speech-to-text for local languages - **Hybrid AI pipeline** – combining LLM reasoning with deterministic safety rules NurseNaija uses AI as a **decision-support tool**, not a replacement for healthcare professionals. ## 4. High-Level Architecture 1. User provides symptoms via **voice or text** 2. Language selection and detection & speech-to-text normalization 3. Hybrid triage pipeline: - Rule-based risk heuristics (offline-safe) - N-ATLaS contextual reasoning - Safety enforcement (no diagnosis, no pr …