With the prominence of machine learning being established in the software market today, special flavors of DevOps must be used to meet the unique requirements of model development, models tuning, and model storage. These key factors can be collected into a philosophy of MLOps and can be empowered by the same technology used in vanilla DevOps. The talk will discuss the different requirements needed for each part of the productization lifecycle of the ML solutions and how they can be deployed in existing DevOps environments.