6G-ROTAM
Resource-optimised training of large language models in 6G networks for AI on Device: application to sensitive medical data
Together with RWTH Aachen, we researched how large language models (LLMs) can be trained in a federated way – that is, without centrally storing sensitive patient data – directly on the device („AI on Device"). A privacy-compliant trust layer secures the communication; the use case comes from psychotherapy. VIVAM GmbH emerged from this first project phase in 2025.
Project at BMFTR →
