Hiring a full-stack developer
A hundred and forty applicants for a title that describes at least four different jobs. The problem here is not that the pile is big — it is that half of it is answering a question you did not ask.
"Full-stack" is a claim, not a description. The title covers a frontend engineer who can write an API endpoint, a backend engineer who can wire up a form, and a genuine generalist who is properly deep in neither — and all three write the identical CV, because the identical CV is what job ads ask for. Worse, the claim is nearly free: touching both halves once qualifies you to say it. So the pile arrives large and miscalibrated, and the useful signal — which half they reach for when it is hard, and how honest they are about the other — is exactly what the format suppresses.
Which half actually matters here
Selene refuses to take "full-stack" at face value, because the honest version of this role always leans one way.
Where the hard problems live
Almost every full-stack role is really 70/30, and the manager knows which way even when the job ad does not say. Getting that on the record early is what stops the guide rewarding people who are shallow in both directions.
One sentence changes the guide
It is an unusual answer and a genuinely useful one, because it inverts a signal most screening treats as good. Breadth claimed evenly across a CV normally reads as strength; this manager has learned it reads as unexamined — and a guide built on his experience ranks the pile very differently from one built on the job ad.
Exceptional, not just good —
“Honestly? Someone who admits which half they’re weaker at. The ones who claim both equally always turn out to be weaker at both.”
→ Honest scoping became a scored criterion; even technology spread stopped counting as a positive.
What the guide ends up measuring
Six criteria on a deliberate 70/30 split rather than an even one. If the manager had said the hard problems were in the interface, four of these six would invert — which is the argument for building the guide from a conversation rather than from the title.
No gates — and the usual proxy is worse than none
Nothing here is a yes-or-no fact, so nothing goes on the application page. The proxy teams reach for instead is a years-of-experience floor, which on this title is close to meaningless: three years of genuinely owning both halves beats eight years of adjacency, and a number cannot tell the difference.
Scored on the split, not the claim
Selene never receives the candidate’s name or the raw CV. She scores a blind, structured profile, with any stated age, sex, nationality and religion stripped out before scoring runs.
What she looks for is where the depth actually sits, which is almost never where the CV’s summary line says it is.
How the blind scoring works →What stays with you
Selene's half
- Getting the real 70/30 split on the record before any criteria are written
- Reading all 140 against the same guide, blind, with the evidence attached
- A ranked shortlist that separates genuine breadth from claimed breadth
- An interview brief that names the half each candidate did not evidence
Yours
- Getting developers to apply — Selene doesn’t post to job boards or source candidates
- Any technical exercise; she reads what people wrote about their work, not their code
- The interview itself, with her brief in hand
- The offer, and every call that matters
Admit which half is harder
Say where the difficult problems actually live and the guide stops being a list of technologies.