How do you hire senior engineers when every CV is perfect?

I recently ran a senior engineering round for a fintech lending platform I work with. Two senior seats. One of them took somewhere north of sixty interviews to fill. Not sixty applications. Sixty conversations. A few years ago that number would have meant something was wrong with the role or the pay. Today it just means the filters that used to do the early work for you are gone.

The CV is dead as a signal

AI writes a flawless CV now. It writes a flawless cover letter too, tuned to your job ad, echoing your own language back at you. Every applicant arrives looking like the exact person you described. Which means the document tells you almost nothing, because the effort it used to represent has dropped to zero.

The screening work the CV used to do has moved into conversation. Someone has to actually talk to these people, early, before anyone senior burns an hour on them. That is a real cost, and most companies have not budgeted for it. They are still running a process designed for a world where a strong CV meant a strong candidate.

The bluff is rational, so expect it

Here is the part that surprised people when I described this round. Candidates will interview well, talk fluently about systems they have supposedly built, push all the way through to the technical stage, and then fail to write a straightforward SQL query. Not a tricky one. A basic one.

I stopped being annoyed by this once I understood the incentive. From the candidate's side, bluffing is close to free. Worst case, they lose an hour. Best case, they land a senior salary at a company that never checked. The bluff will keep coming as long as it keeps working somewhere.

The fix is not outrage, it is process. Verification has to move earlier and get cheaper. A short, practical technical screen, before the long interviews, filters out the fluent talkers at low cost to everyone. If a candidate objects to demonstrating basic competence for a senior role, that is also information.

Job ads gave us volume. Recruiters gave us candidates.

We ran the usual job ads first. They produced a flood, and the flood was the problem, because every application in it looked identical for the reasons above. So we moved the search to recruiters and to direct approaches.

Recruiters worked, but only after I learned the lesson properly: a recruiter is exactly as good as the spec you hand them. A vague brief comes back as a stream of confident wrong candidates, and you will spend your saved time interviewing them. The spec that worked was sharp to the point of being blunt. The actual stack, in detail. What senior means for this seat specifically, not in general. The deal-breakers, stated as deal-breakers. What the first few months of the job actually look like. Once the recruiters had that, the quality of what came back changed completely.

The Rockefeller rule

There is a story about John D. Rockefeller I think about during every hiring round. When he could not find a sound use for investors' capital, he sent it back rather than deploy it badly. He would rather turn money away than lose it.

A senior seat is the same decision. The pressure to fill it is enormous, because every week it sits empty the team feels it. But a wrong senior hire costs far more than an empty chair. You pay the salary for months while you find out. You pay again when the work gets redone. You pay a third time in what it does to the standards of the engineers around them. And then you pay the human cost of unwinding it, which nobody enjoys and everybody postpones.

So the round takes as long as it takes. Sixty-odd interviews for one seat felt absurd in the middle of it. It felt like discipline at the end.

Who actually runs this?

Here is the quiet reason so many mis-hires happen. In most growing businesses, nobody in the building can run a technical interview. HR can assess the person but not the engineering. The founder can assess drive but cannot tell whether the system design answer was real or recited. So the decision falls back on confidence and rapport, which are exactly the qualities the bluffers have practised most.

Someone technical has to sit inside the loop. Someone who has built the kind of systems the candidate claims to have built, and can tell in twenty minutes whether the claims hold. If that person does not exist in your company, borrow one before you hire, not after. It is a fraction of the cost of getting the seat wrong.

That is part of the work I do for the businesses I advise, and it is the part where the damage prevented is easiest to see.