Interview Intelligence Engine reference profile.
What did this candidate demonstrate in the interview, and how do they compare against the role's evaluation criteria?
Overview
Reduces inconsistent, memory-dependent interviewer scoring by grounding evaluations in what was actually said.
Who owns this decision
VP Talent Acquisition, Head of Recruiting
What this Specialist Model actually does
Decision: What did this candidate demonstrate in the interview, and how do they compare against the role's evaluation criteria?
The final taxonomy records 4 supporting or external components, but it doesn't promote an Oumi-owned specialist model for this engine.
Who uses this Specialist Model
- VP Talent Acquisition
- Head of Recruiting
- Chief Human Resources Officer
- Hiring Manager (as end user)
Where the specialist model sits
The final taxonomy records 4 supporting or external components, but it doesn't promote an Oumi-owned specialist model for this engine.
Speech-to-text transcription pipeline
Supporting component
Speech-to-text transcription pipeline for recorded interviews
How Oumi connects to your stack →
Bias and consistency audit layer across interviewers and candidates
Supporting component
How Oumi connects to your stack →
Long-document reasoning LLM
Supporting component
Long-document reasoning LLM for structured summary and evidence extraction against the rubric
Tabular/calibration model
Supporting component
Tabular/calibration model for suggested-score alignment with historical hiring outcomes
How Oumi connects to your stack →
How to evaluate and improve it
Evaluation
- Summary accuracy vs. human-reviewed transcript
- Hiring manager agreement rate with suggested scoring
- New-hire performance correlation with interview scoring
Feedback signals
- Hiring manager agreement or override of the suggested score
- Candidate hired or rejected outcome
- New-hire performance following interviews the engine scored
- Interviewer feedback on summary accuracy


