How Do You Get AI Into Production?
Developer Voices
00:00
How to Monitor Production and Deploy ML Models
When do you know what's when you need to run an other cycle, right? When do you know? It's a really good question. I thought of it myself. One of the things I've seen worked best with companies that I worked with is if you can compare the expected results with the actual result and then assess accuracy. That would be probably the most cost efficient way to assess the quality of the current model that runs in production. There are more fancy methods like, um, data drift, model drift, business drift but those requires more heavy processing of the data always.
Transcript
Play full episode
