Skip to main content
Kovi uses seniority levels to calibrate question difficulty, expected answer depth, and scoring rubrics for every interview session. Rather than applying a one-size-fits-all evaluation, Kovi adapts its line of questioning based on the years of experience and role scope you define — so a junior engineer is never penalized for not knowing distributed consensus protocols, and a senior architect is never let off the hook with a surface-level answer.

​
Seniority Tiers at a Glance

​
Setting Evaluation Level in the SDK

Pass the evaluation_level parameter in your schedule_interview() call to tell Kovi which tier to apply. The value must exactly match one of the strings in the table above.

​
How Scoring Adapts

Kovi uses the same 0–10 scoring scale and the same pass_score threshold across all seniority tiers, but the underlying rubric shifts to reflect level-appropriate expectations. When Kovi grades a response, it benchmarks the answer against what a strong candidate at that specific level should demonstrate — not against a universal absolute standard. In practice, this means:
  • A Junior candidate scoring 6.5 has demonstrated solid command of fundamentals and clear communication for their experience level.
  • A Mid-Level to Senior candidate scoring 6.5 has shown they can reason through production problems, optimize code, and integrate systems effectively.
  • A Senior to Architect candidate scoring 6.5 has articulated credible architectural trade-offs, distributed systems thinking, and leadership judgment.
The same numeric score therefore represents meaningfully different demonstrated capability depending on the tier. You can apply a consistent pass threshold across all your roles without inadvertently penalizing less-experienced candidates or lowering the bar for senior hires.
We recommend setting your pass_score between 6.3 and 7.5 for all seniority levels. As a calibration reference, a highly capable candidate benchmarked at approximately 130 IQ typically scores around 6.5 — use this as your baseline when deciding where to set the threshold for a given role.