AI & Simulation
Models that genuinely perform, and simulations whose results are reproducible. I build hybrid systems: a deterministic core that stays computable and verifiable and machine learning where it is actually the better tool.
What you bring me in for
-
You want to use AI, but the results have to be explainable and repeatable.
-
You want to embed AI in the company and don’t yet know where it really pays off.
-
You hold rule-based knowledge that nobody has made computable yet.
-
You want to simulate decisions before they cost money.
What I do
-
AI transformation in the company
Where AI really pays off, where it doesn’t, and in what order.
-
Applied AI and ML architecture
Selecting, adapting and assessing models for one concrete purpose.
-
Fine-tuning open models
Adapted to your own data, without handing it to third parties.
-
Retrieval augmentation
Models that work on your knowledge instead of general knowledge.
-
Simulation engines
Rule-based, deterministic, testable.
-
Graph and network analysis
Making causal chains, feedback loops and instabilities visible.
-
Digital twins
A computable image of the real system.
-
Evaluation and confidence
Measuring how good a model really is instead of claiming it.
-
Training and briefing
So your team can operate and judge the systems themselves.
Got something stuck between domain logic and technology?
Tell me briefly what it’s about. I’m looking forward to hearing from you.