My Own Medic (M.O.M)
AI-powered medical assistant for accessible health guidance using LLMs
An AI medical assistant that understands your health history. Instead of generic web searches that ignore your medications, allergies, and conditions, M.O.M delivers personalized, context-aware health guidance through a simple chat interface.
Selected for presentation at UN Open-Source Week 2025 in New York.
The Problem
Searching “chest pain” online returns millions of results—from heart attacks to muscle strain. None of them know you’re on blood thinners, allergic to aspirin, or had a stent placed last year. Generic health advice can be dangerous.
M.O.M solves this by loading your complete medical profile into every conversation, ensuring responses account for your unique situation.
What It Does
Context-Aware Medical Chat — Ask health questions and get responses that consider your medications, allergies, diagnoses, and lab results. The AI won’t suggest ibuprofen if you’re on blood thinners.
Conversation Summaries — One click generates a structured summary of your chat, perfect for sharing with your doctor at your next appointment.
Risk Assessments — Based on your conversation and health data, the system flags potential short-term and long-term health considerations.
Real-World Use Cases
Rural Health Clinics — In areas with limited access to specialists, M.O.M can help patients understand their conditions and prepare better questions for their next telemedicine appointment. A patient managing diabetes can ask about medication interactions before their quarterly check-in.
Elderly Care — Seniors often take multiple medications and struggle to remember contraindications. A caregiver can use M.O.M to quickly check if a new over-the-counter pain reliever is safe given the patient’s existing prescriptions, without waiting for pharmacy callbacks.
How It Works
The system uses a 32-billion parameter LLM (Qwen 2.5) with custom prompt engineering. When you ask a question, your complete medical profile—demographics, medications, diagnoses, allergies, recent lab results—gets injected into the prompt context. The AI response is grounded in your specific health situation.
Tech Stack: Python • PHP • JavaScript • Groq API • Qwen 2.5 LLM
HIPAA Compliance
The project incorporates data compliance measures including HIPAA considerations for handling sensitive medical information. Patient data is processed locally and never stored on external servers.