The wrong first technical hire costs you a year — not just their salary, but the months you spend before admitting it isn't working, plus the architecture decisions you now live with. The stakes on this one hire are higher than almost anything else you'll do early.
Most founders screen for the wrong thing: can this person code? At this stage that's table stakes. The question that actually matters is whether they can make a call with incomplete information and own the consequences — because that's 80% of what a first senior engineer does.
Screen for judgment, not trivia
Skip the algorithm puzzles. Put a real, messy decision from your actual backlog in front of them and watch how they think. Do they ask what the business needs before reaching for a tool? Do they scope the smallest thing that works, or gold-plate?
The strongest signal is a candidate who, mid-interview, talks you out of something. That's the behaviour you want on payroll. The one who agrees with everything will build whatever you say — including the wrong thing, quietly, for six months.
The reference question that works
Ask past managers one thing: "What did they own, end to end, that would have failed without them?" Vague answers mean a strong contributor, not a leader. For your first senior hire, you need the second one.