This deliverable is a first version of the deliverable named “Edge data processing and service certification”.
The work has focused on identifying the building blocks required to develop a framework for deploying AI-based data processing and sharing modules at edge data sources, considering edge-cloud interactions following MLOps principles. It involves analysing various ML frameworks such as TensorFlow, PyTorch, tinyML, and edgeML, while addressing security and privacy concerns by implementing adaptive edge anonymization or tagging of sensitive data.
At this stage of the project these topics are still quite exploratory; this report focuses thus more on challenges, considered options and possible implementation choices. The last version of this deliverable, which is due M34 (July 2025), will provide all the details about the technical choices and their implementation.
This document, along with all the public deliverables and documents produced by SEDIMARK, can be found in the Publications & Resources section.
