The AI Reckoning / Chapter 20

Hiring Scorecard

Use consistent questions and evidence to evaluate ownership of real systems.

Download printable PDFOne page · Updated October 20, 2026

Rate each dimension with evidence

End-to-end curiosity: asks about users, data, constraints, and business outcomes.
Failure ownership: explains a specific incident, their contribution, and what changed afterward.
Tradeoff fluency: explains the costs and limits of a recommendation as well as its benefits.
Legacy judgment: investigates existing behavior before proposing replacement.
Maintenance experience: can explain how decisions held up after delivery.
AI discipline: describes what they delegate, what they verify, and how they handle uncertainty.
Business communication: explains technical consequences clearly to people outside engineering.

Use a small unfamiliar codebase and a bounded problem. Explain expectations, permitted tools, and available context consistently.
Watch how the candidate forms and tests a hypothesis. Provide reasonable accommodations and assess the work, not speed or interview polish alone.
Use a 1–5 scale with written examples: 1 = little evidence; 3 = adequate evidence for the role; 5 = strong relevant evidence. Calibrate requirements to the actual job.

Candidate / role: …. Evidence and gaps: …. Decision and reasons: ….

Use the printable worksheet to keep your notes and evidence with your team.