DocumentThe scoring rubric

Dimensions13 · weights sum to 100

Range0–1000 · five calibrated tiers

Revisionv3 · 2026

We will not score your time
without showing the math.

On transparency

The full rubric, the modifiers, the calibrated tiers — published in plain sight.

Every listing is read against your resume across thirteen weighted dimensions. The Scoring Agent returns a calibrated 0–1000 fit score with the evidence preserved — direct quotes from your resume and the job description, dimension by dimension. You can audit any verdict. We score nothing in private.

See how it works Join the waitlist

01 — Dimensions

Thirteen graded dimensions. Weighted by what predicts a real hire.

No.DimensionWeightAnchor scores (0 / 5 / 10)

Required hard-skill coverage

18%

0Less than 30% of required skills present in your resume.

560–70% coverage with at least one shipped use.

10At least 95% of required skills present, with depth and evidence.

Skill depth

10%

0Skill listed but no projects or outcomes attached.

5One project where the skill was central.

10Shipped or owned the skill across multiple years and contexts.

Seniority & scope alignment

10%

0Two levels off — IC4 applying for a director seat, or vice versa.

5One level off — workable with the right framing.

10Exact match on level, team size, and operating scope.

Skill recency

0Last used more than five years ago.

5Last used two to three years ago.

10Currently using the skill in your most recent role.

Years of experience vs. ask

0Under 50% of the requested years, or more than twice the ask.

570–90% of the requested years.

10Within ±15% of the requested years.

Domain & industry match

0Unrelated industry — the playbook does not transfer.

5Adjacent — many patterns transfer with translation.

10Prior direct experience in this exact domain.

Stack specificity

0Stack overlap is essentially none.

5Partial overlap — half the stack matches.

10Full stack match across languages, frameworks, and tools.

Role-shape match

0Inverted — IC role asking for what you did as a manager, or reverse.

5Partial — split between IC and managerial work.

10Exact role shape: same balance of doing, deciding, and leading.

Trajectory fit

0Regression — meaningfully below your last role.

5Lateral — same level, different shape.

10Clear next step in scope, seniority, or ownership.

Mission & product affinity

0Zero signal of interest in this kind of product or mission.

5One indirect signal — adjacent product, related side project.

10Explicit prior work in the exact product space.

Company-stage match

0Only ever worked at the opposite stage (e.g., FAANG into seed).

5Mixed history with at least one similar stage in the past.

10Direct match on company stage and operating constraints.

Logistics fit

0Hard blocker — visa, clearance, location, or remote policy mismatch.

5Partial friction — workable with relocation, sponsorship, or hybrid days.

10Clean fit on all logistical constraints the JD lists.

Differentiator presence

0No preferred-pluses from the JD present in your resume.

5One preferred-plus present and demonstrable.

10Multiple rare pluses present — the hire-on-paper case.

Weights are tuned to predict real outcomes, not vanity matches. Each dimension is scored 0.0–10.0 against the anchors above; the weighted sum is multiplied by ten to produce a final 0–1000 score. Every dimension preserves the exact resume and JD quote that justified the score.

02 — Modifiers

A handful of corrections that override the math when the math is wrong.

M.00 — Hard disqualifier

The override that respects reality.

−200 cap When the listing requires something you cannot offer — citizenship-restricted clearance, country-locked remote, mandatory in-office in a city you do not live in — the verdict is capped below 200, regardless of how well the rest of the rubric scores. Filtering for these early saves you the hour.

M.01 — Compensation alignment

When the JD discloses pay.

±10 to ±25 Bands within your target range bump the score; clearly-below bands move it down. Silent JDs are not penalised.

M.02 — Recency of role exit

When your last role ended yesterday — or two years ago.

±10 to ±20 Active job seekers and recent exits are weighted slightly differently than candidates with a long gap. Honest, not punitive.

M.03 — Stretch-shape adjustment

When you are reaching, deliberately.

±10 to ±15 A modest correction so that a clear stretch role does not get scored against you for being one. Calibration, not consolation.

M.04 — Re-entry signal

When you are coming back to something.

+10 Returning to a stack, domain, or role shape you previously excelled in but recently stepped away from gets a small bump.

M.05 — Conflict tag

When the JD asks for two things at odds.

±10 JDs that demand a research-quality engineer who also ships features weekly get scored carefully — the agent flags the conflict in the verdict text.

M.06 — Geography softener

When relocation is on the table.

+10 to +15 If you have marked yourself open to a region, the logistics dimension is corrected upward for roles in that region.

M.07 — Title-noise filter

When titles lie.

±10 "Senior" at a 5,000-person company means something different than at a 30-person one. The agent reads the scope under the title, not just the title.

03 — Calibrated tiers

A score is just a number. The verdict is the point.

900–1000Exceptional fitStrong match on the heaviest dimensions, no disqualifiers, several differentiators. Worth a personal outreach note alongside the application.

800–899Strong fitA confident application. Coverage and depth are present; tailoring will get this through the ATS and onto a recruiter's desk.

690–799Plausible fitWorth applying. Some gaps to address in the tailored resume, or one dimension where you are reaching. Honest, not hopeless.

500–689StretchA meaningful gap exists. Apply if the company, mission, or role shape is exceptional — but understand the math.

0–499Mismatch · disqualifiedThe rubric does not see a credible path. Either a hard disqualifier applies, or the weighted dimensions do not clear the threshold. Skip and conserve.

04 — What you get back

The verdict comes with its evidence attached.

Returned per scored role

verdictstrong · plausible · stretch · mismatch · disqualified — a one-word label you can sort and filter by.

scoreThe 0–1000 integer from the weighted sum, post-modifiers.

dimensions[ ]All thirteen dimensions with the 0–10 score, the weight applied, and direct evidence quoted from both your resume and the job description.

modifiers_applied[ ]Each modifier that fired, with its delta and the reason it fired. Nothing happens off the books.

top_strengthsThree short phrases for why this role fits — useful for the cover letter or outreach note.

top_gapsThree short phrases for what you would be reaching on. Used to direct the resume tailoring.

summaryA single sentence describing the fit. The line you would read first, written like a person.

Apply to fewer roles.
Apply to the right ones.

Join the waitlist Back to overviewPrivate beta. We invite people in weekly, in small numbers.

We will not score your timewithout showing the math.