Description
We are seeking a talented AI/ML Engineer to develop a lightweight, functional Proof of Concept (POC) for a Medical AI Agent. The goal of this micro-project is to build a scoped Python script or agentic workflow that processes clinical text (such as anonymized patient notes or medical charts) and extracts structured medical insights. This is a well-defined, milestone-based project. Successful delivery of this POC will lead to a larger, higher-budget contract to build out the full platform. Key Responsibilities Structured Medical Extraction: Build an LLM-backed pipeline to parse unstructured clinical text and extract key entities, including diagnoses, prescribed medications, dosages, and active symptoms. Standardized Output: Ensure the agent maps extracted data into a clean, predictable JSON schema (e.g., matching semi-structured formats like synthetic FHIR-lite resources). Agentic Logic/Tooling: Implement basic validation or routing logic (e.g., flag potential drug-to-drug interactions using a mock lookup dictionary or external medical API). Clean Delivery: Provide modular, well-documented Python code that can be easily run and verified locally. Technical Requirements Language & Stack: Python 3.10+ using LangGraph, LangChain, or direct structured outputs via OpenAI/Anthropic APIs. Medical AI Awareness: Familiarity with parsing medical jargon and ensuring strict adherence to deterministic JSON structures (e.g., utilizing Pydantic for schema enforcement). Privacy Mindset: Unders