alb257/afyatech-precision-prompting

Precision prompt engineering for maternal health AI in rural East Africa using AIM, MAP, and safety frameworks. # afyatech-precision-prompting Precision prompt engineering for maternal health AI in rural East Africa using AIM, MAP, and safety frameworks. # 🌿 AfyaTech Precision Prompting – Maternal Health AI ## 📌 Project Overview AfyaTech is a health tech initiative focused on building an SMS-based maternal health assistant for expectant mothers in rural Kenya and Uganda. This project demonstrates how **precision prompt engineering** improves AI reliability, cultural relevance, and safety in low-resource healthcare environments. Generic AI prompts often fail in rural contexts by ignoring: - Local foods (e.g., ugali, matooke, sukuma wiki) - Distance to clinics (>5km for most users) - Limited transport and M-Pesa constraints - Cultural pregnancy practices This repository applies **AIM, MAP, Chain-of-Thought, Verifier, and OCEAN frameworks** to redesign maternal health prompts. --- ## 🧠 Frameworks Used - **AIM (Actor, Input, Mission)** → Defines role, context, and objective - **MAP (Memory, Assets, Prompt)** → Grounds responses in local data and resources - **Chain-of-Thought + Verifier** → Improves safety in medical triage - **OCEAN Framework** → Ensures data integrity and factual correctness --- ## 📂 Repository Structure --- ## 📍 Key Use Cases ### 1. Nutrition Advice Localized dietary guidance using affordable foods like: - Ugali - Matooke - Sukuma wiki - Beans and omena ### 2. Appointment Reminders Context-aware reminders considering: - Travel time (walking/boda boda) - Clinic schedules - Community health worker (CHW) coordination - M-Pesa transport constraints ### 3. Emergency Triage Safe symptom assessment system that: - Asks clarifying questions first - Classifies severity - Prevents panic-driven recommendations --- ## 🎯 Key Insight Precision prompting transforms AI from a generic text generator into a **context-aware healthcare assistant** capable of supporting vulnerable populations in low-resource settings. --- ## 📊 Impact Goal Improve maternal health …