The knowledge layer

Your firm’s intelligence — versioned like an engineering drawing

Vincere holds your records. This holds your judgment— what a good machinist looks like, which clocks matter, when a brief isn’t worth working. Every placement sharpens it. One of those two things is replaceable.

The sealed baseline

Before the software, it was this.The day your real scorecards and process knowledge land, version zero is sealed — signed, dated, immutable. Everything after is a tracked change with a who and a why. Your firm’s know-how stops living in individual heads and starts having a history.

Sandbox edition sealed 28 JUL 2026 · built from trade research + the discovery call

The ladder — nothing gets promoted because someone said so

hunchin developmenthouse standardcalibrateda scorecard is calibrated when placements agree with it — not when anyone asserts it
CNC Machinist (Miller / Turner)house standardN=14 · LAST RUN THIS WEEK

The tier ladder is everything: operate < set < prove-out < edit at the control < program from scratch < program offline in CAM. 'CNC machinist' spans £26k-£58k on that one axis. Named controls, mill vs lathe, sector pedigree and materials tell you in 30 seconds which tier is real.

3 knockouts · 8 weighted criteria14 shortlists scored; the 5-axis gap criterion has predicted the placement 3 times running.
Coded Welder (TIG 141 / MIG 135 / MMA 111)house standardN=11 · LAST RUN THIS WEEK

'Coded' means a current test certificate to a named standard — BS EN ISO 9606-1, ASME IX, or AWS D1.1 — specific to process, material group, thickness and position. 6G pipe is the premium ticket. The clock matters more than the paper: ISO 9606 needs 6-monthly signed confirmation and 3-yearly revalidation; a lapsed code is a retest cost, not a qualification.

3 knockouts · 8 weighted criteriaCoding-clock knockout caught 2 lapsed tickets before submission.
MOT Tester (Class 4 / 7)calibratedN=22 · LAST RUN TODAY

Every tester is 6-7 years in the making: Level 3 LVMR, four years' full-time experience, the DVSA-approved course, a station demo test — then an annual assessment plus 3 hours' recorded training every year (training year ends 31 March). Miss the annual and testing status suspends. The workforce is aging and the pipeline is thin; a current card is the scarcest single credential in the aftermarket.

5 knockouts · 6 weighted criteriaHighest-volume scorecard; annual-assessment knockout has never been overridden.
NDT Inspector — Level II (US)in developmentN=6 · LAST RUN THIS WEEK

The trap everyone outside the trade misses: under SNT-TC-1A and NAS-410, NDT certification is employer-held — a Level II dies the day the inspector leaves. What transfers is the documented hour log and the training record. So you score re-certifiability, not the paper: methods, logged hours per method, written-practice pedigree, vision exam currency. 30% of the certified workforce is over 55; two to three openings per available tech.

4 knockouts · 8 weighted criteriaUS written-practice weighting still being tuned against CT placements.
Quality Engineer (AS9100)in developmentN=8 · LAST RUN 2 DAYS AGO

The standards line is the fastest read: ISO 9001 is baseline; AS9100D means FAIR to AS9102, traceability, concessions, NADCAP exposure — the South Coast filter. IATF 16949 transfers in with a volume-to-traceability culture shift. Then verbs: 8D closures owned, FAIRs authored, CMM programmed vs operated vs 'familiar' — three different people.

3 knockouts · 8 weighted criteriaAS9100 knockout is strict on purpose — 1 strong fit in 35 mirrors the real market.

Sandbox numbers — with real data, N and the run history come from your actual placements

The lead-score recipe — versioned like everything else

A lead score isn’t a magic number and it isn’t fixed. It’s five separate reads, each worth up to 20 points, each answering a question a good desk manager would ask out loud:

DemandAre they buying right now? Live jobs and buying signals.Job records + the signal engine
RelationshipHow warm are we? Terms status and the furthest-along contact.Terms of business + contact stages
InventoryCould we actually supply them? Matched candidates in the pool for their trades.The pool, scored by archetype + geography
LookalikeDo they resemble accounts we've already won? Same sector, size, patch.Your own won accounts — the pattern library
HistoryHave we billed them before? A proven door beats a cold one.Placement history

Freshness is confidence, not points

A stale record doesn’t make an account worse — it makes the read less trustworthy. So age never moves the score; it moves the confidence badge. “65, low confidence, nobody’s called since March” is an honest sentence. “41” pretending to be certain is not.

The recipe changes — with evidence

Weights start from research, then move when your placements disagree with them. If lookalike keeps predicting wins better than demand does, lookalike earns more weight — and the change is dated, attributed, and reversible. Same ladder as everything else on this page.

Recipe v1 · sandbox weights from industry research · calibration against your placements is phase 2

Knowledge that isn’t used goes grey

A scorecard nobody has run in four months desaturates and raises a question: dead knowledge, or dead instrumentation? Either way, somebody should know. Confidence always carries a sample size — small-N claims stay visibly provisional.

What’s yours, what travels, what never leaves

Compounds for you

Your archetypes, your client map, your placement outcomes, your calibrated judgment — the asset this page describes. It appreciates with every placement.

Travels as method

The shape of how this works — scoring engines, clocks, the pipeline mechanics. That’s ours, refined by every build.

Never leaves your vault

Candidate data, client records, transcripts, terms. Structurally fenced — not a policy, an architecture.

And if we ever part ways: you keep the mechanic, not just the car — the knowledge layer is yours, exportable, with its full history.

IN YOUR BUILD · PHASE 2
The living version. Scorecards editable by your team with propose-and-confirm, every change attributed; placements automatically feeding N; the monthly diff — what changed in how you work — generated, never typed.

Everything dashed in this app is a decision, not a promise — you choose what gets built next.