Clinical Intake System

MedTech | An AI agent to streamline healthcare patient intake processes

  • Developed a chat agent providing accessible and personalized support for psychiatric-related queries
  • Developed LLM-powered chatbots using LangChain with FastAPI backend for scalable API services.
  • Integrated audio I/O support and implemented short-term memory (STM) persistence using PostgreSQL.
  • Built long-term memory (LTM) modules for storing and retrieving patient profiles through RAG.
  • Designed agentic RAG pipelines with LangGraph, Chroma embeddings, and PostgreSQL for data persistence.
  • Implemented summarization workflows using Map-Reduce technique for efficient context handling.
Project Description:

Clinical Intake System is an conversational AI system designed to streamline healthcare patient intake processes through intelligent automation. The platform features two specialized AI agents: a Creator Agent that conducts structured clinical interviews to systematically collect patient health profiles across 15+ medical categories (demographics, symptoms, allergies, medications, medical history, etc.), and an Explorer Agent that enables healthcare providers to query and retrieve patient information through natural language conversations.