🏥 EHR De-identifier — AI-Powered HIPAA-Compliant Data De-identification
A powerful EHR (Electronic Health Record) de-identification tool that removes Protected Health Information (PHI) using a dual-layer approach: configurable regex pattern matching followed by LLM-powered analysis — with full audit logging and validation reporting.
Key Features: • Dual-layer PHI detection (regex + local LLM analysis via Gemma 4) • Supports HL7 FHIR, CDA, and free-text clinical notes • Full audit trail with before/after comparison reports • REST API via FastAPI + CLI interface • Docker-ready deployment • 100% local processing — no patient data ever leaves your machine
Tech Stack: Python 3.10+, Ollama (Gemma 4), FastAPI, Docker, ChromaDB
Built by Nrk Raju Guthikonda as part of a 116+ open-source AI tools portfolio. GitHub: https://github.com/kennedyraju55/ehr-deidentifier
Log in or sign up for Devpost to join the conversation.