MediMind: Leveraging Docker Microservices and Gemini Pro for High-Availability Clinical Decision Support

Abstract

This demonstrative document details the development, integration, and implementation of MediMind, a web application hosted on the cloud and powered by AI which utilizes Google’s Gemini 2.0 Pro Model for health information retrieval and medical image analysis. The system provides a conversational interface for health queries and both automated image analysis during food expert systems and exercise evaluation, as well as automated pill recognition. We describe the system architecture focusing on the components that incorporate AI models and newer computing paradigms based on containerization and cloud deployment in modern ecosystems. This paper discusses the steps taken to containerize the system using Docker, outline deployment options on Google Cloud Platform, including Cloud Run (serverless) and virtual machine, and analyze security concerns, scalability, and performance for medical care services in the cloud. Our results proved the application of generative AI and cloud-natives SDK architectures offers framework agility and ease of use for builders which could be helpful for enabling advanced personal healthcare systems, nutrition and fitness counseling, and even pedagogy in medicine.