Michael Falk

Built and debugged internal .NET/C# administrative tools for Epic-related workflows, adding audit-friendly steps and reducing repeated manual checks during release windows. Investigated production integration defects using SQL and HL7/FHIR-style payload comparisons, authored Selenium/Postman regression coverage, and implemented Git/Azure DevOps change-controlled configuration management.

Built a retrieval-augmented generation pipeline for clinical documentation to reduce hallucinated recommendations and improve citation traceability for reviewers. Redesigned LLM inference APIs with token budgeting, batching, and Redis caching to control runtime cost and keep latency predictable during peak usage.

Delivered managed-services support for healthcare clients by troubleshooting Epic EHR workflow configuration, validating REST/SOAP payloads, and isolating failures to specific mapping rules and interface contracts. Implemented change-controlled EHR configuration updates with SIT/UAT verification, audited ETL steps to improve data reliability, and supported SLA-focused ticket workflows in Jira/Serv

Developed applied NLP and enterprise knowledge extraction workflows, including RAG-style retrieval services to reduce manual search effort and improve answer relevance. Fine-tuned and evaluated LLM components with PyTorch and added monitoring and regression checks to detect quality drift before releases.

Served as a primary point for Epic workflow build and troubleshooting, translating client requests into build/test plans while preserving clinical workflow integrity. Led configuration audits and change-controlled Epic migrations, improved incident triage documentation, and strengthened integration stability by debugging HL7/FHIR-like payload validation and mapping discrepancies using Postman and

Earned a Bachelor's degree in Computer Science from Georgia Institute of Technology, graduating in May 2015.