Credentials are issued by Talent2Role, a product of Quantum Task AI (Dubai, United Arab Emirates), from the Cyber edition at cyber.talent2role.com. Contact: info@quantumtaskai.com. Quantum Task AI is a Dubai AI Seal enterprise, a signatory of the EU AI Pact and the CSA AI Trustworthy Pledge, and listed on the CSA STAR Registry (Level 1 self-assessment, CAIQ v4 answers public); those are statements about how we operate as an issuer, not endorsements of any credential or holder.
| Track | What it means | What it does not mean |
|---|---|---|
| Evidence-aligned | The holder's self-reported evidence (résumé text, roles held, skills, projects) was analysed by the Talent2Role engine against a NICE Framework work role and the result is stated as coverage and a readiness score with a confidence band. | No examination was taken. The evidence is the holder's own account. Treat it as a structured statement of what the holder claims. |
| Knowledge-verified | In addition, the holder passed a timed, randomised knowledge check for the work role (item bank published per role, pass mark set by standard-setting) within the last 12 months. | The sitting is online and unproctored. It demonstrates knowledge of the role's competency statements, not professional experience or fitness for a particular job. |
Every credential carries its track on the certificate face, on the verify page and in the registry record, together with: the work role and its NICE ID, coverage (matched statements ÷ total), the readiness score and confidence band, the career stage the score was measured against, the method versions (engine, rule set, ontology, mapping prompt, item bank), whether AI-assisted extraction was used, any consistency flags, and the issue date.
A credential exists only once it is registered. The application sends the claims to the registry, which assigns the ID, signs the record and stores it; only then is the certificate drawn. The printed ID and the QR code resolve to /v/<id>, which renders the registry's copy — the link itself carries no data.
T2R-XXXX-XXXX-X. The alphabet excludes 0/O/1/I/L and the final symbol is a check character, so a mistyped ID is rejected before lookup.signature.fields (keys sorted recursively, no whitespace). The signature is base64url(r‖s)./.well-known/t2r-issuer.json with their kid. Retired keys stay published so older credentials keep verifying.sha256$…, Open Badges 3.0 style), never as the address itself.OpenBadgeCredential JSON-LD document and as a compact JWS signed with the same key (/api/creds?id=…&format=ob3 and …&format=ob3jwt)./api/creds?id=<id>, build the canonical JSON of signature.fields, and verify signature.sig against the key whose kid matches. The verify page does exactly this in your browser and shows the result.A credential may be revoked if the evidence behind it is found to be false, the holder breaches the honour statement, a knowledge sitting is found to have been taken by someone else, or the issuer discovers a defect in the method version it was issued under. Revocation marks the registry record revoked with a date and reason; the verify page and the Open Badges status endpoint show it. The signature remains valid — a revoked credential is a genuine credential that has been withdrawn, not a forgery. Holders are told by e-mail and may appeal to info@quantumtaskai.com.
Evidence-aligned credentials are point-in-time statements dated at issue. The knowledge-verified track is valid for 12 months from the passed sitting; the Open Badges document carries a matching validUntil. Re-assessment is free.
Before the first credential is issued, the holder confirms that the evidence is their own and accurate, that they are 18 or older, and that they sit any knowledge check themselves without assistance. Sittings are timed and randomised; tab switches and paste attempts are counted and recorded on the result (a flagged sitting is marked provisional). Consistency checks on the evidence (for example advanced-tier claims at an early career stage) mark a credential provisional on its face and in the registry. The full policy is in Trust → Assessment integrity.
The scoring method is published at /method.html. Every record carries the versions it was computed with, so a score can be reproduced. Changes to the engine, rule set, item banks or mapping prompt bump a version; credentials issued under earlier versions remain valid and say which version they used.
Practitioner reviewers of the mapping rubric and item banks are listed here as they join. Reviewers lend their judgement to the method; they do not vouch for individual holders.
Quantum Task AI · Dubai, United Arab Emirates · info@quantumtaskai.com · quantumtaskai.com