Immutable attempt history
Every attempt retains its scenario version and timestamps.
The Thrive With AI Talent Hub evaluation standard
Readiness must be earned through evidence.
“Day 1 Ready” is intended to mean that a candidate has met a versioned, role-specific verification policy. It is not a course badge, guarantee of performance, or promise of placement.
Current launch status: the versioned scenario, evidence submission, attributed human review, integrity, and policy-decision workflow is operational in early access. Automated checks are recorded from external sandbox runs until the isolated execution worker is connected.
What is evaluated
Engineering judgment under realistic constraints.
Scenarios test the ability to design, build, evaluate, explain, and improve AI systems, not merely call an LLM API.
Problem framing and requirements
Architecture and engineering judgment
Implementation quality and testing
Evaluation and observability
Reliability and failure recovery
Security and responsible AI
Deployment and operations
Documentation and communication
How a decision is formed
Multiple checks. One attributable decision.
- 01
Versioned scenario
The candidate receives fixed requirements, deliverables, constraints, and deadlines.
- 02
Automated checks
Safe, isolated checks examine functionality, quality, grounding, security, and deployment health.
- 03
Human review
An assigned evaluator scores observable evidence against a structured rubric.
- 04
Policy decision
Verification is granted only when all active policy conditions are satisfied.
Verification policy
No single score can override a critical failure.
The active policy is configurable and versioned. A candidate must satisfy every required condition, not only an average threshold.
- Required scenario completed
- Automated checks completed
- Human review completed
- Overall weighted threshold met
- Critical-dimension minimums met
- Security and reliability requirements passed
- Required evidence present
- No unresolved integrity flags
What employers can access
- Verification state, date, validity, track, and level
- Skills supported by the verified scenario
- Candidate-published project evidence
- Role-fit criteria, evidence sources, and explicit gaps
- Project impact, technology, and evidence level
Hidden tests, private scoring guidance, confidential scenario criteria, integrity details, and private reviewer notes are not exposed.
Conflict prevention
Evaluators cannot verify their own submissions.
Time-bound validity
Verification expires under the active policy and can require reassessment.
Audited overrides
Manual decisions, revocation, and overrides require an attributable reason.
Build your profile while the evaluation layer is prepared.