In order to make Machine Learning suitable for the masses and accessible to a wide range of users and applications, as well as to bring real added value, its development must be cost-effective.

An ecosystem that wants to achieve this must remove the currently most important limitations in AI development:

• Lower the current entry barriers to ML, through a flexible, holistic and structured ML development environment.
• Promote network effects for deep collaboration within a team and across an entire community in order to centralize and facilitate the exchange of technical and content-related ML elements (e.g. models, data and the like).
• Increasing work efficiency with a focus on iterative, structured and embedded workflows to improve the quality of data and ML models.
• Ensure full reproducibility of the entire value chain to promote confidence and quality of the applicable ML model.

We at MLReef have defined these points as our goals. 

MLReef is an MLOps platform for efficient, collaborative and replicable work on Machine Learning (ML) projects. It is globally the first to set its core element on community collaboration to enable instant reuse of any ML element previously published on MLReef. This drastically shortens development time and increases the model quality through fast and structured iterations.

USP
This unique selling proposition means that user generated public data, AI models, data visualization and data processing can be used immediately by everyone in a single environment. This network effect accelerates the innovative power of our target groups enormously, since ML projects do not always have to be developed from scratch but can build on existing projects very easily and quickly.

Target market
MLreef is designed to work for a broad range of customer segments, from academia, individual data scientists, small teams (2 – 3 individuals) and enterprises.

Space connection
Data gathered by ESA´s Earth Observatory (EO) is an ideal base for many scientific and commercial ML applications. In this context, MLReef provides a short bridge to a holistic ML environment to create downstream research and commercial application. The aim at ESA BIC is to define and develop technical data integrations with leading EO based data providers (EODC, TU Vienna, DIAS).