Using the GrowthBook platform to manage ML experiments

Within the product development, depending on a company's stage of maturity, different ways of conducting the product hypothesis cycle can be found:

• some use all sorts of "crutches" to separate user groups on the backend or frontend, and product managers manage the process;
• some prefer to consolidate everything within the development team, which manages the entire test;
• others allocate a separate team to develop their own SDK or implement proprietary/free tools for test management.

Each of these approaches is possible, as it allows you to optimize a particular development metric at the expense of something else: for example, speeding up TTM in exchange for an increased rate of technical debt accumulation. In this talk, we want to talk about the way to organize experiment pipeline, where the responsibility for launching and testing features lies within the ML development team, based on the open-source GrowthBook platform. The proposed approach is intended to reduce the number of integrations on the side of the core development team, while increasing the speed of bringing new versions of machine learning models into production.