Hire Zechen Y. - AI Engineer - New York

MCIT, Division of Applied AI technology Research of GPT information extraction stability on unstructured clinical notes, guiding application of LLM in clinical setting.

* Developed an AI clinical decision support application, user select or trigger specific diagnosis workflow (in demand of real-world doctors and patients) on the interface. (eg. CHF endpoint prediction within five years) * Distributed agents across multiple servers via API, enabling a scalable system architecture suited for commercial * applications. * Enhanced the capability of intelligent multiagent workflow by real-time server-to-server agent communication to * Enabled retrieval augmented generation. Clinical notes, AHA guidelines and UMLS three-layer medical knowledge graph with emphasis on temporal connection. Reducing hallucination and providing transparency on reasoning of how agent * arrives to the diagnosis.