Hiring Rubric

Hiring Rubric

Before you start

Gather the following from the user:

1. What role are you hiring for? (Title, level, team)
2. What does success look like at 6 months? (3-5 concrete outcomes the hire should achieve)
3. What are the must-have vs nice-to-have skills? (Technical and non-technical)
4. How many interview rounds? (Phone screen, technical, system design, behavioral, culture)
5. Who is on the interview panel? (Names and which dimensions they'll evaluate)

If the user says "we need a strong engineer," push back: "Strong at what? Backend systems? Cross-team collaboration? Debugging production issues? Define 3-4 specific capabilities that matter most for this role."

Hiring rubric template

1. Role Profile

Write a concise profile that anchors the rubric to actual job needs, not a generic job description. Include: role title, team, level, reporting line, and 2-3 concrete 6-month goals the hire should achieve.

2. Evaluation Dimensions

Define 4-6 dimensions. Each dimension must be independent — avoid overlap. Assign a weight reflecting its importance to the role.

| Dimension              | Weight | Assessed In          |
|------------------------|--------|----------------------|
| Technical depth        | 30%    | Technical interview  |
| System design          | 25%    | Design interview     |
| Problem-solving        | 20%    | Technical interview  |
| Communication          | 15%    | All rounds           |
| Collaboration          | 10%    | Behavioral interview |

3. Scoring Criteria

For each dimension, define what a 1 through 4 looks like. Avoid vague language — describe observable behaviors.

Dimension: Technical Depth

4 - Strong Hire:  Solves the problem correctly with clean, production-quality
                  code. Identifies edge cases proactively. Discusses tradeoffs
                  of their approach without prompting.

3 - Hire:         Solves the problem with minor issues. Handles most edge
                  cases when prompted. Can articulate why they chose their
                  approach.

2 - Weak:         Reaches a partial solution with significant guidance.
                  Misses important edge cases. Struggles to compare
                  alternative approaches.

1 - No Hire:      Cannot make meaningful progress on the problem. Shows
                  gaps in fundamentals expected at this level.

Write a scoring rubric like this for every dimension. Use concrete behaviors, not personality traits.

4. Interview Questions

For each dimension, provide 2-3 questions with follow-ups. Behavioral questions must use the "Tell me about a time..." format to elicit past behavior, not hypotheticals.

For technical dimensions, use role-specific coding or debugging problems with follow-ups like "How would you test this in production?" For behavioral dimensions, use "Tell me about a time..." questions that elicit past behavior with follow-ups exploring outcomes and lessons learned.

5. Scorecard

Create a standardized scorecard every interviewer fills out within 24 hours of the interview.

Candidate: _______________    Interviewer: _______________
Role: _______________         Date: _______________
Round: _______________

| Dimension              | Score (1-4) | Evidence (required)           |
|------------------------|-------------|-------------------------------|
| Technical depth        |             |                               |
| System design          |             |                               |
| Problem-solving        |             |                               |
| Communication          |             |                               |
| Collaboration          |             |                               |

Overall recommendation:  [ ] Strong Hire  [ ] Hire  [ ] Weak  [ ] No Hire

Key strengths:
Key concerns:

The "Evidence" column is mandatory. A score without a specific observation is not valid.

6. Calibration Process

Before interviews begin, have all interviewers score the same mock interview independently, then compare. Align on what a "3" looks like for each dimension with a concrete example. During the process, interviewers submit scorecards before the debrief — the most junior interviewer presents first to prevent anchoring.

Quality checklist

Before using the rubric, verify:

• [ ] Every dimension maps to a real job requirement, not a generic trait
• [ ] Scoring criteria describe observable behaviors at each level
• [ ] No two dimensions overlap significantly (test: could one score high on dimension A and low on dimension B?)
• [ ] Behavioral questions ask about past experiences, not hypothetical scenarios
• [ ] The scorecard requires written evidence for every score
• [ ] Weights total 100% and reflect actual role priorities
• [ ] The calibration process is documented and scheduled before interviews begin
• [ ] At least one dimension assesses collaboration or communication

Common mistakes to avoid

• Generic dimensions. "Technical skills" is too broad. "Ability to design fault-tolerant distributed systems" is specific enough to evaluate. Tie every dimension to what the person will actually do in the role.
• Hypothetical interview questions. "What would you do if..." invites rehearsed answers. "Tell me about a time when..." surfaces real behavior. Always prefer behavioral questions for non-technical dimensions.
• Missing the evidence requirement. Without mandatory evidence, scorecards become gut-feel ratings. Require interviewers to write the specific moment or statement that justified their score.
• Anchoring in debriefs. If the hiring manager shares their opinion first, everyone adjusts toward it. Always have the most junior interviewer present first, and submit scores before the meeting.
• Overweighting technical skills. A brilliant engineer who can't communicate or collaborate will slow the team down. Ensure at least 20-25% of the weight covers non-technical dimensions.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: The-AI-Directory-Company
• Source: The-AI-Directory-Company/agents-and-skills
• License: MIT
• Homepage: https://ai-directory.company

Install and usage instructions live in the source repository linked above.