The latest articles on DEV Community by Pyreddy Bhagirath Reddy (@bhagirath_reddy_3d5ddb418). https://dev.to/bhagirath_reddy_3d5ddb418 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4149257%2Ff2e16ebf-70e6-44fc-b054-628fb9e8c1f9.jpeg https://dev.to/bhagirath_reddy_3d5ddb418 en Pyreddy Bhagirath Reddy Tue, 29 Sep 2026 09:49:32 +0000 https://dev.to/bhagirath_reddy_3d5ddb418/why-i-spent-more-time-on-fake-data-than-real-code-4anp https://dev.to/bhagirath_reddy_3d5ddb418/why-i-spent-more-time-on-fake-data-than-real-code-4anp <p>Before a single line of extraction logic existed for LungTrace, I spent hours writing fictional patients. That felt backwards at the time. It turned out to be the highest-leverage work in the whole project.</p> <h2> The demo is only as convincing as the data </h2> <p>LungTrace tracks a patient's lung scan findings over time using <a href="https://github.com/vectorize-io/hindsight" rel="noopener noreferrer">Hindsight</a> as the memory layer, and flags follow-up scans that were recommended but never happened. The idea is easy to explain. Proving it works is not, unless the data is built to prove it.</p> <p>If I'd generated 20 random synthetic reports, the odds that any of them would show a clean growth pattern, a genuinely missed follow-up, and a stable control case were low. So I stopped treating the data as filler and started treating it as the test suite.</p> <h2> Designing patients like test cases </h2> <p>Every patient I wrote had a job to do:</p> <ul> <li>A clear growth case: A nodule at 3mm, then 4.5mm, then 6mm across three reports spanning eight months. Read alone, each report says "small nodule, likely benign." Read together, it's doubled in size.</li> <li>A missed follow-up case: One report, a recommendation for a follow-up scan, and a due date that's now well in the past with nothing after it.</li> <li>A stable, boring case: Same size across three annual scans. This one mattered as much as the alarming ones, because a system that flags everything is as useless as one that flags nothing.</li> <li>A shrinking, reassuring case: A nodule getting smaller, correctly marked as needing no further follow-up.</li> <li>A different-organ case: A liver lesion instead of a lung nodule, to check the grouping logic didn't secretly assume "lung" everywhere.</li> </ul> <p>Each report also needed to read like something a radiologist actually wrote, not a data table in paragraph form, since the whole point was testing whether an LLM could pull structure out of prose:</p> testing ai python datascience