What the number is made of

Methodology

01What this scores

The scorer reads a public LinkedIn profile the way a startup hiring manager screens one: employers, scope, timing, what's verifiable. It measures the profile, not the person. A strong engineer with a thin profile scores low. Change the profile and the score changes.

02The data

A read fetches the public profile the moment you run it. Every employer on it is looked up for headcount, funding and stage as of the date the person joined, so a company is judged as it was then. GitHub is fetched and graded only when the profile links it.

03Five measured categories

Employers and scopeWhich companies, and how much they owned there

Startup fitEvidence they have worked somewhere small and early

Job tenureHow long they stay, and whether they move up while there

EducationWhether the school is a serious one for computer science

Public workCode and projects anyone can go and look at

Listed heaviest first: employers weigh most, public work least. Each category is scored 0 to 100 on the rules in its footnote on the read page. A category with no evidence is skipped and its weight splits across the others, so a missing section never scores zero.

Education stands in for evidence, so it steps back as evidence arrives. The stronger the employment record, the less the school counts, and at the top it counts for nothing. By then the work has answered the question.

04The overall judgment

An AI model scores the full profile against a fixed written rubric. The model version is pinned, so every profile is scored by the same system. It also weighs what the five categories leave out: location (a hub city reads in-person ready, a bare country reads remote-only, outside the US raises timezone and visa questions for the call), trajectory, and whether the career holds together.

The top of the scale isn't flat. Clearing a hard hiring bar puts a profile in the exceptional band; the last few points go to people who built or run the thing itself, work another engineer in the field would recognize without the logo attached. Building from nothing counts wherever it happened, including inside a big company: founding a product that reached millions is the same evidence as founding a company.

Scoring runs at fixed settings but isn't perfectly deterministic: re-run the same profile and the score can move a point or two.

05The number

The five category scores and the overall judgment combine at fixed proportions, with the judgment carrying more. That blend is lifted onto the display scale, which runs 0 to 99. There's no 100: a read of a page is never the whole person, so the top stays open.

One method governs every score. Change the method and everything it already ranked is re-scored on the new one, so the leaderboard and a fresh read always speak the same language. A score is a reading taken on a date.

Rank and percentile compare the result with every profile scored so far. The percentile counts scores strictly below yours. Tied scores share the better rank.

96 to 99exceptional

87 to 95strong

80 to 86solid

61 to 79emerging

0 to 60unproven

The leaderboard is working engineers only: running a public figure through the scorer gives them a card, not a seat. Ties are ordered by votes. Votes are compared only after scores, so no amount of voting moves anyone past a higher number.

06The holds

Three rules bind after the judgment, on the numbers, because a model told a rule in prose will bend it for a famous name. Each fires on what the record actually lists; a section the page leaves out never triggers one.

Education at 60Schools listed with no degree, or degrees with none in computer science or an adjacent field: the education category is held at 60. Any CS-adjacent degree, at any level, clears it. The other four categories are untouched.

Read at 74No hands-on engineering role anywhere on the page holds the whole read at 74, beneath the solid line. Founder, executive and advisory titles aren't building. The card says so where the tier word goes: Held at 74, no hands-on engineering role on the profile. Linus Torvalds, Demis Hassabis and Jensen Huang all hold here.

Startup fitA founding role floors the startup-fit category at 85. A history entirely inside large companies caps it at 78.

A hold is about the page, not the person. Add the role and it lifts on the next read.

07The characters

Every read is also one of twelve characters, taken from the shape of the five categories and the titles on the page. It says what kind of engineer the page describes. No points, no effect on rank. The card, the share image and the leaderboard all draw the same one.

Zero to OneBuilds the thing before there is a thing
In the OpenThe work is public and the record is the proof
OperatorTakes the early thing and makes it run
PathfinderDoes best before the process arrives
AnchorStays long enough to see the second-order effects
MainstayFinishes what gets started
SpecialistGoes deep where most people stop
ArchitectDesigns the system other people build in
Blue ChipHas shipped inside rooms that set the bar
ResearcherComes at it from the theory first
ScholarTrained where the hard problems are set
BuilderShips the work the product actually runs on

08What a read can't see

A read only knows what the profile shows it, so private repositories, work under NDA and everything an interview would turn up are all invisible to it. It also picks up the same habits as the screen it is imitating. A well-known employer counts for more than it probably should, being in a hub city counts for more than it probably should, and someone who spells out what they owned will beat someone who did the same work and never wrote it down. That is a real bias rather than a rounding error, and every read says so on the page.

09Zelcomp, the pay estimate

Some reads come with a pay band we call Zelcomp, and it is not a guess about the person. It is what employers certified to the US Department of Labor for the kind of role the profile describes, drawn from about 519,000 H-1B wage filings across FY 2024 and FY 2025. Filings that sit exactly at the prevailing wage are dropped first, since that number is a legal minimum rather than a real offer, and dropping them takes out most of the IT-services filings. The rest are grouped into cells by level, role family and metro, then by the employer itself or by companies of a similar size once there are enough filings to support it. A cell needs twenty filings behind it before it will answer, and the read tells you which cell it used.

Certified wages run behind the market, because a company files once a year and posts jobs continuously, and the gap is widest at the bottom of the band. That is why posted ranges are read alongside them: roughly four thousand salary bands from the public careers pages of about three hundred companies, each folded in at its midpoint. They cover the startups that are too small to file anything, and they pull both ends of the band toward what is being offered now instead of what was certified last year.

Tested against employers the cells had never seen, the middle half of the band catches 49 percent of filings and the wide band 78 percent, with a median error of 14 percent, or 12 when the employer has its own cell. Checked cell by cell against BLS, Indeed, Levels.fyi base figures and public H-1B trackers, the employer cells land within a few percent, and so do the SF and Seattle level cells. Startups, the AI labs above all, advertise above what they certify, so a posted range can run 20 to 30 percent over this band. The total-comp line widens the base band by what stock and bonus usually add at an employer that size, and it stays deliberately wide. A bare title says nothing about seniority, so the highest rank on the record and the years behind it move the band up the pool it matched, and the panel says when that happened. Engineering leadership comes with a caveat: at that level most of the money is equity and negotiation, so the band is a floor, sometimes priced from the staff pool, and the label says so. Founders set their own pay, so their band reads as what the role pays employees. Roles outside engineering get no band.

Score your profile