The latest articles on DEV Community by Ines Mota (@ines_mota). https://dev.to/ines_mota 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%2F4215044%2F76004b6f-e70c-4642-bdfa-616c14c83f39.jpg https://dev.to/ines_mota en Ines Mota Thu, 29 Oct 2026 10:47:42 +0000 https://dev.to/ines_mota/the-boring-half-of-shipping-a-model-doj https://dev.to/ines_mota/the-boring-half-of-shipping-a-model-doj <p>The model was the easy part. It took two weeks to train something good enough, and about five more to make it something the team could rely on.</p> <p>The half nobody puts in the writeup:</p> <p>Data validation before inference. If a field arrives empty or in a new format, the model still returns a confident answer. We added a schema check and a rule: if validation fails, return a clear error instead of a guess.</p> <p>A version pair. Every prediction logs the model version and the preprocessing version, because they drift independently and a mismatch produces results that look plausible.</p> <p>An eval set that lives in the repo and runs in CI. Three hundred labeled examples, run on every change. It has caught two 'harmless' preprocessing tweaks that quietly cost four points.</p> <p>A cost number. Not accuracy, not F1, just the price per thousand documents. It shapes the design more than any other metric.</p> <p>None of this is interesting at a conference. It is the difference between a demo and a service.</p> machinelearning mlops python