The project PREDICT-6G, where Eviden led T3.4 on Frictionless inter-domain orchestration and network control, ended successfully on June 2025 after 2.5 years of work involving colleagues from the MLOps team. To close it, a Final Review was organized online by the EC on September 30.

The consortium presented the outcomes of the project by performing five demos focused on showcasing: the operation of both Multi-technology multi-domain data-plane (MDP) and AI-driven Multi-stakeholder Inter-domain Control-Plane (AICP), the integration of the MLOps Framework and WiFi Digital Twin to deploy a Federated Learning model, and the use of deterministic communications in a Smart Manufacturing use case. 

The main results produced by Eviden in the project have been the following:

• Extension of the MLOps Framework to support Federated Learning in deterministic environments.
• Integration of the MLOps Framework with a WiFi Digital Twin to develop and serve a Federated Learning model devoted to set the best traffic policy in distributed WiFi domains.
• Development of Service Automation module as part of AICP design

Since the project has achieved all the expected results, the outcome of the final review was very positive. Eviden’s contribution has been key to the project by: (i) developing and integrating the MLOps solution that allowed to introduce the predictability component as part of the AICP design, and (ii) developing the Service Automation module used to orchestrate the AICP operations needed to provision and decommission E2E and Technology specific services.

This project aligns with the roadmap of the MLOps team towards the extension of the MLOps framework to support distributed learning architectures.