analysis
Learning the domain before building for it
Taking on problems in fields I did not start in, and learning the field well enough to make decisions inside it rather than only implementing someone else's.
It shows up in the choices nobody asks you to justify: which threshold, which metadata matters, which part of a paper process was essential and which was an artefact of paper. Get those wrong and the software runs correctly and is still useless.
Evidence3
Case-tracking system for a legal department
The requirement came to me as a description of a paper process rather than a specification. Most of the work was learning how the office actually runs a case, so that I could tell which parts of that process were essential and which were only artefacts of doing it on paper.
Microscopy image processing and 3D reconstruction
Fluorescence microscopy was new to me. Before I could write anything useful I had to learn how CZI stacks and their metadata are organised, what the DAPI and GFP channels hold, and enough about iso-levels to settle on triangle thresholding for the surface extraction.
Certhub: a six-person team build with a real client
Angular was the team's choice rather than mine, so picking it up fast enough to carry the same load as everyone else was part of the work.