AI & LLM research from Tuningen.
We work on methods, tools and models for sovereign AI. Publications, preprints and collaborations — as soon as they are publicly citable, they appear here.
Publications appear here once they are publicly citable.
We publish our research only after it is peer-reviewed or available as a preprint with a public DOI. Once that's the case, this page will list title, authors, abstract, venue, DOI, PDF, citation and optional project links — structured and machine-readable.
Research areas
- / 01
Sovereign LLM infrastructure
Efficient inference stacks for on-premise deployments with clearly documented data paths.
- / 02
Domain-specific adapters
LoRA and MoE architectures for SME domains — industry, government, legal, healthcare.
- / 03
Eval methodology
Reproducible evaluation of domain-specific LLM applications with hybrid metric suites.
- / 04
Secure AI integration
Architecture patterns and audit pipelines for integrating AI into existing enterprise software.
Research collaborations are welcome.
If you work at a university, Fraunhofer institute or a research consortium and you find our topics interesting, we'd love to talk.