Fifteen years making systems reliable, now applied to production AI.

Production AI needs the same reliability rigor that safety-critical engineering brings. That conviction runs through my work, and it is why I describe what I do as a rare combination rather than a career change.

For fifteen years I have built complex systems to be safe and dependable in regulated environments. Today that means automotive systems engineering at Audi and CARIAD, the Volkswagen Group’s software company, including functional safety under ISO 26262 and ASPICE, where a defect is a recall and not a bug ticket. I trained as a Dipl.-Ing. in mechatronics at TU Dresden and added deep-learning and MLOps training, at DeepLearning.AI on Coursera and in AI/ML nanodegrees at Udacity, to point that engineering discipline at applied AI. The full record is in my CV and on LinkedIn.

I have also stood on the other side. Several of the systems I have built ran inside an operating business, where I was the person answerable for whether a rollout survived contact with real users. Since then I treat adoption as a behavioural question and not only a technical one. The hardest part of a system is usually getting people to rely on it.

I own systems from the data model to deployment, so the architecture stays coherent end to end. Cost is a design criterion: much of practical AI engineering is knowing where not to call the model. In everything I build, AI is a powerful component and never the whole product.