Building CPI’s capability in industrial connectivity, Federated Learning and data science methodologies

The purpose of this tender is to demonstrate CPI’s capability in data analytics and Federated Learning by supplying an analytics-enabling data infrastructure for a plastics recycling rig (Lot 1), and a complete Federated Learning platform (Lot 2), initially orientated towards furnace applications which is deployable at different sites/locations. These lots are to be supplemented by a training programme CPI has identified in support of these activities (Lot 3).

This tender is split into three lots where are described in detail in the Technical Requirements section and accompanying URB’s. The intended outcome from this lot is a Federated Learning platform that allows machine learning (ML) modelling of data from industrial assets and plants.

The platform should allow building of ML models on the edge without the need to share data. The intended outcome for this lot is the provision of training in data scientific approaches.

The target audience of the training are general scientists (typically chemists / chemical engineers, trained to PhD level or graduates with several years’ experience) with good understanding of data science and python coding, but no formal training. The focus of the training should be on practical approaches and robust procedures that can be implemented at CPI.