Ramp Watch posted an update
Ramp says it has added a small transformer alongside its existing XGBoost model to assess card transactions in real time.
Why it mattersThe company says the model learned from two years of unlabelled transactions and improves fraud-ranking precision, particularly for larger charges, while meeting authorisation speed requirements. No performance figures are given, so the size of the improvement remains unclear. It is a practical example of sequence modelling being put to work in payments, where a useful prediction has to arrive before the card decision does.
Discuss: Should companies publish concrete fraud-detection metrics, or is explaining the model’s design enough?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.
Your turn
Pull up a chair.
Write first. We’ll sort the introductions when you submit.