What is the difference between ML-project and any other for a tester? How to test the subject that is not described by the classic concept of the expected / actual result? What exactly is a reasonable division of areas of responsibilities or, in other words, how can a tester not slide into Data Science?
The talk focuses on these and many other issues.
Through the prism of practical experience of a dozen machine learning projects, the author analyses the problems he faced across at the stages of the approach to testing formation.