The Fourth Industrial revolution brings the implementation of such technologies like Big Data, Internet of Things, Virtual Reality, Augmented Reality, Machine Learning and Artificial Intelligence. Current market requirements are changing rapidly and use of machine learning algorithms based solutions can significantly increase business competitive advantages in the context of globalization. Back in 1995 on “Neural Information Processing Systems Workshops” an idea of creating a platform that could provide a reliability of data sources and trained neural networks has been declared. Platform implementation would lead to the rapid development of the whole stack of machine learning based technologies and help to reduce the cost of its development, mass adoption and significant growth of derived systems efficiency. The concept of rationality of several machine learning models merging with their further transfer learning has been proposed and proved later. The common limiting factor of development and implementation of similar systems was the lack of reliable technology that could provide a decentralised digital trustworthiness for final machine learning models and data sources.