OS4ML: Open Space for Machine Learning

Abstract

Currently, artificial intelligence (AI) and machine learning (ML) are central in many discussions. Unfortunately, their use is predominantly limited to specialists in the field. To broaden the application of AI, the barriers to its usage need to be substantially reduced. The challenge for companies planning to use AI lies in the need to hire qualified experts and invest in expensive, powerful hardware. These issues are addressed in the Open Space for Machine Learning (OS4ML) platform, that is developed in this project. The open-source solution emphasizes user-friendliness, enabling domain experts without AI technical skills to apply machine learning to their data. This method eliminates the necessity for costly and time-consuming individual AI projects. The platform is based on Kubernetes and uses a microservices architecture, ensuring flexibility and scalability. It incorporates various powerful open-source tools, setting a standard for scalable AI applications. A key step in democratizing AI is the transition from low-code ML tools to nocode tools. This involves the development of a user-friendly frontend, which makes AI more accessible to a wider audience by removing the need for extensive coding knowledge. The platform is designed to be adaptable to any cloud environment and is easy to set up by the community. This strategy leverages cloud computing benefits, such as scalability and cost efficiency, and provides the option for users to host the platform on their premises, a crucial feature for handling sensitive data.