William Lee

Ai Automation Engineer @Jorie AI

US

OpenCV / Python / Selenium / OCR / RCM Process Automation / MSSQL / Azure / LLMAutomated core components of the RCM process for healthcare insurance clients—covering eligibility checks, claim edits, and status inquiries. • Built a Python-based automation system that processes claims daily, cutting 16+ hours of manual work to under 90 minutes. • Interpreted PDDs, resolved logic gaps, and translated them into automation rules. • Applied OOP to centralize utilities (logging, error handling, note posting, validations). • Integrated Selenium, SQL, and JSON rules for resilient EHR navigation and field validation. • Used GPT-4 and Copilot to generate helper code, then reviewed/refactored outputs for production. • Logged feedback on LLM results (edge cases, logic flaws, maintainability). • Built agent-based task queues for scalable, multi-bot execution via modular Python workflows. • Automatically generated Excel/SQL reports with result flags, screenshots, and remarks.OCR & Document UnderstandingUsed Azure OCR to extract diagnoses, procedures, and provider info from scanned visit notes. Supported RAG-based pipelines to classify billing data into standardized categories.Multi-Client DeploymentsDeployed tailored RCM automations: • Client A: OpenCV + OCR for image-based form extraction • Client B: Selenium for insurance eligibility + claim status automation • Client C: Browser automation for claim editsThese cut manual workloads and ensured RCM compliance.End-to-End Automation StackUsed Selenium (browser automation), Azure OCR (document parsing), MSSQL (data storage), Bitwarden + ScreenConnect (credential and VM management) for secure, persistent automation.Tech Stack:Python, OOP, Selenium, Azure OCR, OpenCV, MSSQL, Bitwarden, VM, Git, Dev Box, Linux/Unix scripting