video1.0<iframe src="https://www.loom.com/embed/5eddc60609344c8c8cf8b6827b3e800d" frameborder="0" width="1152" height="864" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>8641152Loomhttps://www.loom.com8641152https://cdn.loom.com/sessions/thumbnails/5eddc60609344c8c8cf8b6827b3e800d-014d1fe09871d61c.gif265.715Hi everyone, I built a mid cloud clawback AI to stop Medicaid billing scams by bridging messy whistleblower tips with massive government datasets. When an investigator uploads a tip, my NVIDIA and IAM powered system extracts text, providers, and specific medical billing codes, then extracts and verifies the target codes and flags anomalies. It cross references targets against a 3GB national Medicaid dataset with a high speed DuckDB engine, and I show mathematical proof via Z scores, like a provider at 15,000 standard deviations above the monthly average for a simple office visit. Finally, one click produces an FCA compliant ready file with statistics. You are asked to upload the sample Visible.pdf and click Extract and Verify to see the output.