frontier research deep techai governancealgorithmic fairnesseu ai actmodel evaluation
Complete evaluation framework
What to assess and how to score it
Review the evidence signals before interviewing. Then use the anchored descriptions—not instinct alone—to choose the score that best matches each answer.
01
Evaluation factor
Theoretical command
35% weight
Probe command of fairness metrics (demographic parity, equalised odds), value alignment literature, and instruments like the EU AI Act, NIST AI RMF, and OECD principles.
Evidence to listen for
Explains the underlying theory at the level the role demands, and can go a layer deeper when pushed
Knows which results are established and which are contested
Distinguishes their own contribution from the field's
Comfortable saying where the theory runs out
Five-point scoring guide
1
Poor
Recites terminology without understanding; cannot go one layer deeper.
2
Needs Improvement
Surface familiarity; conflates established results with speculation.
3
Satisfactory
Solid grasp of the core theory; thin at the frontier.
4
Very Good
Strong command; separates settled results from open questions.
5
Excellent
Names specific fairness definitions, explains where they mathematically conflict, and cites concrete AI Act risk tiers or RMF functions accurately.
02
Evaluation factor
From theory to hardware or code
30% weight
Ask what they built or shipped: model cards, red-team protocols, bias audits in Fairlearn or AIF360, dataset documentation, impact assessments tied to a real model release.
Evidence to listen for
Has built, simulated, or run something real, not only published about it
Knows the gap between the idealised model and the actual apparatus or system
Names the practical constraint that dominates in real conditions
Can describe a result that did not match prediction
Five-point scoring guide
1
Poor
Purely theoretical; no contact with implementation.
2
Needs Improvement
Some exposure but unaware of practical constraints.
3
Satisfactory
Has implemented work; understands the main real-world limits.
4
Very Good
Strong practical record; articulate about theory-versus-reality gaps.
5
Excellent
Shows artefacts they authored, such as a completed FRIA, eval suite, or audit that changed a launch decision or model card.
03
Evaluation factor
Research judgement
20% weight
Test how they choose which harms to chase when evidence is thin: sampling a harm taxonomy, scoping red-teaming, deciding when a system should not be deployed.
Evidence to listen for
Chooses problems by tractability and value, not novelty alone
Knows when to abandon a line of work
Reads and evaluates others' results critically
Can say what would falsify their own approach
Five-point scoring guide
1
Poor
Chases novelty; no sense of tractability or when to stop.
2
Needs Improvement
Weak problem selection; persists past the point of value.
3
Satisfactory
Reasonable judgement within a defined programme.
4
Very Good
Selects problems well and knows when to abandon a line.
5
Excellent
Prioritises harms by severity and reversibility, admits open questions, and describes a case where they recommended halting or restricting deployment.
04
Evaluation factor
Explaining it to non-specialists
15% weight
Judge whether they can move engineers, legal counsel, and executives without jargon: policy memos, ethics review boards, disclosure language for users and regulators.
Evidence to listen for
Explains the work to an engineer, an executive, or a funder without either mystifying or dumbing it down
Writes clearly
Collaborates across disciplines
Makes the case for resources in terms the audience cares about
Five-point scoring guide
1
Poor
Cannot communicate outside their specialism.
2
Needs Improvement
Explanation is either impenetrable or hollow.
3
Satisfactory
Adequate with technical peers; less effective with lay audiences.
4
Very Good
Explains clearly to specialists and non-specialists alike.
5
Excellent
Translates alignment or bias findings into decisions and trade-offs a product lead grasps, with evidence a recommendation was adopted.
Put this rubric to work
Score every candidate against the same standard
Add these weighted factors to Hirevire and let AI evaluate recorded answers against your rubric.
