In the scientific sense. I do not stop at "it works." I want to know *why* it works, under what conditions it stops working, and what the failure mode looks like. That shows up in the talks, in the physics publications, in the blog posts on compiler behaviour and memory models. It's the deepest consistent thread across my entire career.
I produce knowledge, not just software. The 100+ blog posts, the three public talks, the 15+ papers, the open-source repositories — these aren't side hobbies. They're evidence that I process what I learn and give it back in usable form. Most engineers consume; I also convert.
My choices — topics, domains, the problems I write about publicly — are considered rather than opportunistic. I didn't give talks on trendy subjects. I gave talks on floating-point arithmetic and CI/CD pipelines because working engineers need those things.