Aaish Faisal Hameedi

AI pipeline transforming medical documents into ICD-10/CPT codes. Engineered a Graph-RAG-based retrieval strategy that fused semantic search, knowledge graph links, and deterministic rule lookups, reducing unnecessary prompt context by 30 to 40% and strengthening explainability for code selection. Improved LLM cost efficiency by 25 to 35% through chunk-level retrieval, specialty scoped prompts, and selective context assembly, avoiding repeated full-document inference across multi-document submissions.

Multi-Agent RCM System · Healthcare AI · US Market

An enterprise-grade multi-agent AI system for medical billing and revenue cycle management (RCM). A LangGraph Controller Agent coordinates 6 specialized GPT-5 sub-agents — from clinical document ingestion through ICD-10/CPT/HCPCS coding, human-in-the-loop review, and autonomous claim drafting. Coded 380 medical codes from 63 clinical documents in validation, every code linked to supporting text from the original note.