I blanked for eight seconds in a staff engineer interview after leaning on AI to prep - and it failed me. Here's what I should've done instead, and why I keep getting pulled toward the role that bridges technical depth and business acumen.
Around 2021, if you wanted AI in a data workflow, you picked a vertical. Coding assistance. Data validation. Each was isolated. That constraint is gone now — and it changes everything.
Code comments, unit tests, documentation, diagrams — the stuff that chronically didn't get done. Gen AI changed that completely, and the impact on data teams is bigger than most realize.
I spent years building OLAP cubes. Carefully maintained conformed dimensions, aggregation tables, multidimensional models. That infrastructure was real work. Today, it's increasingly obsolete.
I used to tell junior engineers to study for certifications. I don't say that anymore. Here's the advice I gave a co-op student last month: build something and put it on GitHub.
