Hey everyone,
I just wanted to share our latest blog article.
60% of practical ML projects are stuck in the proof-of-concept to production gap. One of the main reasons for that is data shift. In the article, I describe a framework for how to deal with it.