Tackling Fraud with Tecton

apply(conf) - Apr '21 - 10 minutes

Feature stores enable companies to make the difficult leap from research to production machine learning. At their best, feature stores allow you to code up your features once, then use them for training and production, and share them between models. You can quickly and reliably serve features to your production models so your customers aren’t waiting for predictions.

In this talk, we’ll walk through a common use case for Tecton, developing and fielding a fraud model. We’ve written a blog and included source code on it, so we’ll touch on the high points and save you the good stuff you can dive into at your leisure.

Matt Bleifer

Product Manager


Matt Bleifer is a Product Manager at Tecton focused on core product development for both Tecton and Feast. Prior to joining Tecton as an early employee, Matt was a Product Manager on Twitter’s Cortex team working on ML infrastructure, research, and NLP services. While at Twitter, he helped spearhead the development of the Twitter Feature Store which was adopted company-wide and has greatly accelerated machine learning development. Matt also spent time as a Product Manager at Workday developing applied ML solutions for large enterprises. He holds a BS in Computer Science from Cal Poly, San Luis Obispo.