AI Consulting
From tools to agents: strategy, architecture, and the building of agentic systems — remote, secure, hands-on.
You know the moment: your pipelines run, your servers are automated — and the next question looms. Software that no longer merely executes commands but works on its own behalf changes how infrastructure has to be thought through: agents need clear boundaries, controlled access, and an architecture you can follow.
That is where I work. As a Senior DevOps Engineer and Infrastructure Architect, I help DACH and European companies make the step from classic automation to agentic stacks with clarity — building on open foundations (such as OpenClaw, Hermes) and local LLMs where they pay off. No buzzword bingo, with direct experience from Linux production environments.
Service Pillars
Before any code is written, the architecture is defined: which processes suit agents? What data and access does the stack need? How do governance and security stay architectural building blocks from day one rather than retrofits? I shape your AI decisions into something your engineering team can actually build.
From vision to running system: I design and build agentic stacks on Linux — orchestration, local LLM integration, explicit permission levels, traceable logging. Where appropriate, HETTWER IT takes the company-wide roll-out and Linux-Server-Admin the infrastructure operations — while I stay your personal architect.
Offer Blocks
Can your company actually do something with AI agents? The audit answers that honestly: I assess your setup, processes, and IT environment and show where agents create real leverage — and where they don't (yet). Deliverable: a written report with a prioritized roadmap plus a joint debrief.
Request the Agentic Readiness Audit →For tech leaders and executives: in strategy sessions we develop your agentic roadmap — fitted to your existing automation, compliance requirements, and budget logic. The outcome is an architectural blueprint you can build on for years.
Request Vision sessions →Concrete building after the planning: stack architecture, selection and setup of the orchestration (e.g. OpenClaw, Hermes), integration of local LLMs, clean handover into operations — documented well enough that your team (or our partner network: HETTWER IT, Linux-Server-Admin) runs the stack reliably.
Request a stack project →A system in which AI agents take on tasks on their own: orchestration (e.g. OpenClaw, Hermes), language models (local or hosted), controlled access to data and tools, plus an automation layer. My part: the architecture that makes it safe.
Both, in deliberate order: first clarity and architecture (my core work), then implementation — often via HETTWER IT, and operations via Linux-Server-Admin. You are never left hanging between phases.
Yes, where it pays off. Local LLMs give you data sovereignty and predictable costs; where more capable cloud models are needed, they stay inside a controlled architecture. The audit shows the sensible mix.
Agents receive rights — which is why least-privilege access, traceable logging, and clear boundaries are there from day one. See also my NIS2 work: for me, agentics is an architecture topic, not an experiment.
Sometimes the honest answer is: not (yet). That is exactly what the Agentic Readiness Audit is for — it does not automatically say "yes"; it shows where agents genuinely carry weight.
A free intro call, then the audit or vision sessions, then architecture and build — remote across Europe & DACH, in clearly staged steps.
Tell me in a free 30-minute intro call (video or phone) what your company is working on — I will tell you openly whether agents are the right path for you.