AI Health Assistant

The AI Health Assistant (AHA) is a multimodal system designed to support nurses and healthcare professionals with fast, reliable, and context-aware assistance. Hosted entirely on Google Cloud Platform (GCP), the system integrates advanced large language models, retrieval-augmented generation (RAG), and specialized agents for text, image, and audio to address real clinical challenges.



AHA’s architecture combines a robust backend with an orchestrator, ETL pipelines, and a search engine connected to diverse data stores, including MongoDB, Qdrant, Redis, and Google Cloud Storage. This design allows the assistant to retrieve accurate medical knowledge, process multimodal inputs, and deliver trustworthy outputs in real time. The Text Agent handles Q&A and clinical summarization, the Image Agent interprets medical documents or scans, and the Audio Agent transcribes and analyzes conversations between patients and healthcare providers.

To ensure safety and reliability, AHA leverages GPT for reasoning, Whisper for transcription, and Hugging Face models for domain-specific medical tasks. Evaluation on large health datasets shows that AHA consistently outperforms baseline models in answer relevancy and response stability.

By reducing nurse workload and enhancing decision support, the AI Health Assistant demonstrates how cutting-edge AI can be responsibly applied to improve healthcare efficiency, accessibility, and patient outcomes.