What I’m looking for
This is the reference NEXUS gives GPT-5.6 Luna when evaluating a job. Write naturally—experience to credit, target work, compensation tradeoffs, locations, and true deal-breakers.
Tiering method
Instant job rules
Rules use facts already extracted from each posting. Add, change, disable, or delete one rule without re-running GPT-5.6 Luna. Unknown facts stay neutral.
No structured rules yet. GPT-5.6 Luna’s base tier remains unchanged.
Use this for nuanced career context and experience evidence. Put salary, location, work arrangement, and other changeable constraints in Instant job rules above.
How GPT-5.6 Luna represents priority
These are semantic judgments—not score ranges. Every tier should include an evidence-based reason and explicit uncertainties.
Strong evidence for the core work, acceptable practical tradeoffs, and no true eligibility blocker.
Most core work is supported; remaining gaps are transferable or learnable.
Meaningful overlap exists, but several core requirements or practical concerns remain.
Wrong track, true hard blocker, invalid role, or extreme mismatch.
Current runtime
Selected method: GPT-5.6 Luna semantic tiers. Existing legacy scores remain visible until those jobs are re-tiered; new AI-tiered jobs do not require a numeric score.