Shreya Shankar — Operationalizing Machine Learning
Gradient Dissent: Conversations on AI
00:00
Is There Still Data Corruption?
The challenge is in identifying the corruptions at the time scale that engineers risk react and respond to books. I remember at my first job, all the features for some reason, like there's no technical reason this is necessary. But someone started just like naming each feature column. That's like like four letters, like literally like four letters for no reason. So you don't like put models in production that like won't do well on future data.
Transcript
Play full episode
