Uber's ML Systems (Uber Eats, Customer Support), Declarative Machine Learning - Piero Molino - The Data Scientist Show #064
The Data Scientist Show - Daliana Liu
00:00
The Importance of Alignment in the Evaluation of Models
The biggest example of that was a situation where there was a misalignment in the way the performances were needed to be evaluated. I think that taught me to try to get as much alignment as possible from the very beginning. There were some things that were changed after fact to make it so that our model actually was used for real, but that was a moment of fiction. And so spending more time on evaluation and understanding for real what's to be evaluated is very important.
Transcript
Play full episode
