What I do

I’m a mid‑career engineer based in Sweden with a Master’s in Cognitive Science and roughly a decade of professional work across academic neuroscience, energy‑market risk modelling, and embedded computer vision.

The thread connecting those very different rooms has always been statistical thinking: looking at residuals, watching distributions misbehave, and finding the assumption underneath that quietly breaks. I’m relatively quick at the diagnostic loop: see what a model is doing, guess why, propose a better fit, repeat.

The work I enjoy most starts with a question that hasn’t been answered yet (a sensor that drifts, a signal that hides a structure, or a pipeline held together by undocumented tribal knowledge) and ends with something a colleague or customer can actually use.

Core strengths

  • Statistical and signal reasoning. Time series, image and signal processing (EEG / MEG / MRI and industrial vision), residual diagnostics, model refits.
  • End‑to‑end problem solving. From vague idea to working system: protocol design, data collection, labelling, pipelines, automation.
  • Bridging code and hardware. Pick‑and‑place machines, reflow ovens, smart cameras, e‑paper panels. I like seeing what my code is doing in the physical world.
  • Translating complex systems. Maintainable documentation, clear explanations, and interfaces that survive the engineer who built them.

Tools & domains

My background is deliberately broad, but a handful of technical themes have stayed consistent across roles. In practice that means a mix of modelling, data work, and code that has to talk to physical things.

Technologies

  • Python
  • Matlab
  • C++
  • Lua
  • PyTorch
  • OpenCV
  • MMDetection
  • Ollama / local LLMs
  • Embeddings & RAG
  • Time‑series & risk models
  • Signal & image processing
  • Embedded & industrial vision
  • Raspberry Pi & small Linux

How I work best

I have ADHD that was diagnosed late, a Non‑24 sleep tendency, and a higher‑than‑average sensory baseline. That triangle means energy management matters more to me than to most engineers, and I’ve made my peace with it.

In practice: I do my best work at 30–32 focused hours a week, on a mix of short and medium projects where progress is visible and closure is possible. I’m reliable under real deadlines and burn out fast in perpetual “everything is urgent” modes.

I aim for “good enough for reality” rather than abstract perfection, and I work best with a line of sight to the people who will actually use the thing I’m making.

Collaboration style

Direct, empathetic, low on office politics. I tend to connect easily with other neurodivergent people, and I prefer honest conversation over careful choreography.

I’ve been a team lead and stepped back voluntarily; coordinating others on top of a real technical load isn’t where I add the most value. I’m most useful as a deeply involved individual contributor in a small team.

What I’m drawn to outside work

A few long‑running interests shape what I read, build, and pay attention to. They show up in the projects and essays here in oblique ways.

Long‑running interests

  • AI and cognition. Agent‑like systems, intuitive interfaces, and AI that supports human agency rather than replacing it.
  • Hardware, repair, and reverse engineering. Small tools, e‑paper, batteries, firmware, and the joy of making old things work again.
  • Art and storytelling. AI‑assisted surreal art, impossible spaces, the occasional short story, and slow nonfiction.
  • Nature, flight, and resilience. Hang‑gliding, time outdoors, and a long interest in energy, resources, and how societies adapt to change.

What I’m looking for

I’m interested in roles and collaborations I can stay in for years, without losing sight of why the work matters.

Good fit if you need someone to

  • Apply statistical and signal skills to messy real data.
  • Work close to hardware, sensing, or physical products.
  • Keep a line of sight to actual users or customers.
  • Operate with enough autonomy and trust to do careful work.
  • Contribute to something that won’t feel pointless in ten years.

If this resonates

I’d be happy to talk. The most reliable way is email; I read everything, even when I can’t answer right away.

Email me