Igor Demidov is a machine learning engineer with a conviction that data science code deserves the same care as production software. He writes about clean code practices for machine learning, including why df is a bad name for a dataframe and how intention-revealing names save future readers, and future you, a great deal of pain.

His article on naming variables in machine learning, published on the DataTalksClub blog, applies lessons from software engineering classics to the everyday habits of data scientists. Igor shares his writing and projects on his Medium blog and his GitHub.