We are now part of NAFEMS, an international community of more than 30,000 engineers, designers, and analysts working in engineering modelling and simulation, across aerospace, automotive, energy, chemicals, and process industries. It is not-for-profit, founded in 1983, independent of any software vendor, and for four decades it has been where simulation practice is shared, taught, and pushed forward. Being part of that exchange is a real step for us.
A community asking exactly the questions we build for
AI is moving quickly in engineering simulation, and NAFEMS sits right at the centre of that conversation. Its work on simulation governance asks what it takes to trust a model trained on data: better traceability, validation methods that fit machine learning, and a way to judge whether a training set is good enough. One of its 2026 courses is called "AI Model Validation for Simulation Data: Knowing When to Trust and When to Doubt".
That is one of the central questions STOCHOS was built to address. It learns from the data an engineering team already has, simulation runs and test results, and gives back a prediction together with a confidence interval instead of a single number. When a prediction moves beyond well supported regions of the training data, the reported uncertainty can increase and make that limitation visible. Uncertainty quantification is not an add-on here, it is the method, and it is genuinely exciting to bring that into a community working on the same problem.
What we are looking forward to
- The people. More than fifty NAFEMS events a year, many open to members at no extra cost, and thousands of engineers who work on simulation every day. Conversations there can lead to new perspectives, collaborations, and projects.
- The working groups. This is where guidance on AI and simulation governance is discussed and developed, and where we would like to contribute what we have learned building probabilistic models for small, expensive data.
- The resource centre. More than 8,000 books, presentations, videos, articles, and online training courses, decades of collective experience for our engineers to draw on.
- New collaborations. Across industries and research groups that are asking the same questions we are, and often facing them first.
Not our first NAFEMS moment
The connection goes back further than the membership. In October 2023, ZF Friedrichshafen presented a benchmark of PI-BO, the Bayesian optimization in the STOCHOS package, at the NAFEMS seminar on artificial intelligence and machine learning in CAE-based simulation in Munich. The comparison covered three model variants with 22, 46, and 94 free parameters at three sample set sizes, measured by R squared and MAPE.
It meant a lot that a partner chose to put the method in front of that audience.
Now we are in the room as members, and we are looking forward to everything that comes with it: the events, the working groups, the debates, and the people building the future of engineering simulation.
Learn more: how STOCHOS combines surrogate modeling with uncertainty quantification for reliable engineering predictions.
