AI as an engineering tool, not a slogan
MCS operates its own offline knowledge base for technical documents and uses AI every day in engineering. That supports faster research, stronger drafts and traceable analyses without pushing sensitive customer data outside unnecessarily.
Own knowledge base for technical documents
Our self-developed RAG system runs offline on an NVIDIA DGX workstation and processes technical documents in German, English and Russian. It combines semantic search, keyword search, OCR and structured analysis of tables, manuals and code artifacts.
The three central modes are source-grounded Q&A, specification drafting and code analysis.
AI tools in daily work
Our regular toolkit includes Claude Code, Codex, n8n, Open WebUI and LM Studio. We use cloud models where speed matters, and offline models plus our own RAG system where confidentiality and plant security matter more.
Integration into automation
In plant projects, we use AI where it helps maintenance and availability, for example predictive diagnostics and anomaly detection. Safety-related and legally relevant metering functions remain clearly separated from that layer.