Teaching
Three courses that end with a system you can measure.
Each runs three days and 21 contact hours, and is built around a single codebase that gets progressively harder to fool. Every syllabus below is the one a client receives — including the paragraph stating where the course stops. A three-day course that claims no limits is a course nobody should buy.
- 3 days · 21 contact hours
- Delivered in English or German.
- On-site in Europe, or remote. Wels / Linz, Austria is the home base.
01
LLM Engineering
From prompting and statistical inference to reliable RAG systems
How to build an assistant that answers from your own documents, cites where each claim came from, and ships with a test suite that catches it inventing things.
- Engineers, data scientists, technical leads and advanced AI practitioners.
- An evidence-grounded assistant with hybrid retrieval, citations, tests and safety controls.
02
Reinforcement Learning
From Markov decision processes and control to PPO, continuous control and RLVR
How software learns to make decisions by trying things and being scored on the result — the method behind game-playing systems, robot control, and the way modern reasoning models are trained.
- ML engineers, researchers, simulation engineers, robotics developers and advanced students.
- A small Gymnasium environment with baseline control, trained policy, ablation and multi-seed evaluation.
03
Agentic Engineering
From tool calling and MCP to governed, testable agent systems
How to take an AI agent that works in a demo and give it permissions, tests, an audit trail and a stop button — so you can let it near a real system.
- AI and software engineers, architects, security specialists and technical product owners.
- A governed research or operations agent with typed contracts, an MCP server, evaluation and red-team evidence.
Advisory
The work either side of a course.
Training is usually not the whole problem. Three engagements come up often enough to name.
-
Architecture review
A read of an existing LLM or agent system against the same criteria the courses teach: trust boundaries, evaluation coverage, failure handling and what the system does when it is wrong. Ends with a written decision record, not a slide deck.
-
Evaluation build-out
Standing up the measurement layer a team keeps postponing — golden sets, regression runs, groundedness and trajectory checks wired into CI, so a model or prompt change stops being a leap of faith.
-
Governance evidence
Turning a working system into something an auditor or a board can read: intended purpose, risk register, tool and data inventory, human oversight and a deployment recommendation mapped to NIST AI RMF, ISO/IEC 42001 and the EU AI Act.
Book a training
Tell me which course and roughly when. You get a reply from me, not a sales sequence — dates, a quote, and an honest answer if the course is a poor fit for what your team actually needs.