Dominic Portain

My career has followed an unusual path into applied AI, and I think that path is a strength rather than a detour. I hold a Master of Science in Cognitive Science, and my working skill set spans applied statistics, signal processing, deep learning, classical computer vision, embedded deployment, LLM pipeline design, financial modeling, and technical documentation. Before moving into industry, I spent three years at the Max Planck Institute for Cognitive and Brain Sciences running a research project that involved MRI-based conductivity modeling, EEG/MEG source localization, and large-scale data processing — work that gave me a deep foundation in experimental design and computationally intensive analysis.

My first industry role was as a Senior Energy Analyst at Trianel, a German energy trading company, where I automated the entire reporting infrastructure, trained neural networks for demand prediction, and overhauled the portfolio risk model after identifying that its core statistical assumptions were wrong. I taught myself the relevant financial mathematics, implemented a nonlinear Heston model from scratch, and delivered results that satisfied the risk department — all within a domain I had never worked in before.

I then moved to Sweden and spent five years in industrial AI, first at Woodeye/Microtec and then at SICK IVP. At Woodeye, I built an object detection system for wood defect classification from scratch — starting as a six-month prototype that earned me a permanent position, and eventually scaling into a multi-defect production system. I grew the AI department to five people, managed customer relationships, and handled everything from training data curation to hardware infrastructure. At SICK IVP, I worked on embedded AI for industrial cameras in a properly Agile environment, contributing to firmware in Lua and C++, designing adversarial test datasets to find edge cases, and prototyping forward-looking features like RAG-based documentation search and AI-powered OCR.

The thread that runs through all of this is a repeating pattern: I enter an unfamiliar domain, learn its logic fast, build a working system that solves a real problem, and bridge the gap between the technical implementation and the people who need to understand and trust it. I'm strongest in small-to-medium teams where I can stay close to the customer, hold the whole problem in my head, and move quickly from idea to working prototype to stakeholder presentation — without needing three different people in the room.

I'm looking for applied AI or data engineering roles with high autonomy, task variety, and direct contact with end users. I work best at around 80% time and deliver my strongest results when I can structure my own schedule. In return, you get an unusually cross-disciplinary engineer who can go from customer conversation to trained model to clear documentation in one continuous workflow.