What Dimensions Are Used to Evaluate Job Fit in the AI Job Search Framework?

The AI Job Search framework scores each job posting against five dimensions—Technical Skills Match, Experience Match, Behavioral/Culture Fit, Location & Logistics, and Career Alignment & Motivation—to calculate a weighted fit score that drives the /apply and /rank commands.

The MadsLorentzen/ai-job-search repository implements a structured evaluation system defined in the Job Evaluation Skill. This framework automates candidate-job compatibility assessment through a rubric-based scoring mechanism, transforming subjective gut feelings into objective, comparable metrics.

The Five Evaluation Dimensions

The core scoring logic resides in .claude/skills/job-application-assistant/04-job-evaluation.md. Each dimension serves a distinct role in the assessment pipeline, with four contributing numeric scores and one acting as a binary gate.

Technical Skills Match

This dimension measures the alignment between required and preferred technical competencies with the candidate’s actual capabilities. It produces a numeric score ranging from 0 to 100, directly impacting the final weighted average.

Experience Match

Evaluates how closely the candidate’s work history—including prior functions and industry domains—matches the role’s stated expectations. Like Technical Skills, it generates a 0 to 100 score based on relevance and depth of prior experience.

Behavioral / Culture Fit

Assesses compatibility between the candidate’s behavioral profile and the company’s culture, values, and team dynamics. This dimension outputs a 0 to 100 score reflecting cultural alignment.

Location & Logistics

Functions as a Pass/Fail binary gate rather than a numeric score. This dimension evaluates practical feasibility factors such as commute distance, relocation requirements, and travel expectations. A "Fail" status filters the job from further consideration regardless of other scores, while an optional "Flag" may indicate concerns requiring manual review.

Career Alignment & Motivation

Determines the degree to which the role advances the candidate’s long-term career goals and provides genuine professional energy. It contributes a 0 to 100 score to the final calculation.

Dimension Weighting and Score Calculation

According to lines 34-38 of the evaluation file, the framework applies specific weights to calculate the overall fit score:

• Technical Skills Match: 30%
• Experience Match: 25%
• Behavioral / Culture Fit: 15%
• Career Alignment & Motivation: 30%

Location & Logistics does not contribute to the weighted average; it serves solely as an eligibility gate. The computed weighted average maps to categorical thresholds defined in lines 42-47: Strong, Good, Moderate, Weak, or Poor fit.

Pre-Scoring Eligibility Gates

Before the five-dimensional evaluation executes, the framework applies mandatory pre-scoring filters:

1. Eligibility Gate: Validates basic hard requirements
2. Language Gate: Confirms language compatibility

Only job postings passing both gates proceed to the dimensional scoring phase.

CLI Implementation: /apply and /rank Commands

The evaluation framework operates through two primary CLI commands that invoke the five-dimensional rubric.

Evaluating Individual Jobs with /apply

Implemented in .claude/commands/apply.md, the /apply command processes a single URL or raw job posting text:

claude /apply https://jobindex.dk/job/1234567

Execution flow:

1. Parse the job posting content
2. Execute Eligibility and Language gates
3. Score the posting across all five dimensions
4. Return formatted output showing individual dimension scores and the weighted result

Example output structure includes a table displaying Technical Skills (e.g., 78/100), Experience Match (e.g., 65/100), Behavioral Fit (e.g., 85/100), Location status (PASS), and Career Alignment (e.g., 70/100).

Batch Processing with /rank

The /rank command, defined in .claude/commands/rank.md, batch-scores previously scraped job postings:

claude /rank

This command iterates through the job database, applies the same five-dimensional logic to each entry, and returns a ranked list ordered by weighted fit score. Each entry includes the individual dimension scores or "FAIL"/"FLAG" indicators for pre-scoring gates.

Optional Salary Benchmarking

Lines 24-40 of the evaluation file reference an optional Salary Benchmark feature. When a salary data source is configured, this metric appears in the output but explicitly does not affect the core five-dimension fit scoring or weighted average calculation.

Framework File Architecture

The evaluation system relies on these specific components:

• .claude/skills/job-application-assistant/04-job-evaluation.md: Defines dimension criteria, scoring rubrics, weights, and output formats
• .claude/skills/job-application-assistant/SKILL.md: Registers the "job-application-assistant" skill and command handlers
• .claude/commands/apply.md: Implements single-job evaluation workflows
• .claude/commands/rank.md: Implements batch ranking workflows
• tools/check_framework_version.py: Ensures framework version synchronization when evaluation logic changes

Summary

• Five dimensions drive the evaluation: Technical Skills (0-100), Experience Match (0-100), Behavioral Fit (0-100), Location & Logistics (Pass/Fail), and Career Alignment (0-100)
• Weighted calculation applies 30% Technical, 25% Experience, 15% Behavioral, and 30% Career Alignment; Location acts as a binary gate
• Pre-scoring gates (Eligibility and Language) filter candidates before dimensional scoring occurs
• Threshold categories (Strong, Good, Moderate, Weak, Poor) classify the final weighted average
• /apply and /rank commands execute the evaluation logic against individual or batch job postings

Frequently Asked Questions

How does the AI Job Search Framework calculate the final fit score?

The framework computes a weighted average using four numeric dimensions: Technical Skills Match (30%), Experience Match (25%), Behavioral/Culture Fit (15%), and Career Alignment & Motivation (30%). Location & Logistics serves as a binary pass/fail gate excluded from the weighted calculation. The resulting score maps to categorical thresholds (Strong, Good, Moderate, Weak, Poor) defined in the evaluation file.

What happens if a job fails the Location & Logistics check?

Location & Logistics operates as a mandatory binary gate rather than a scored dimension. If marked "Fail," the job posting is immediately disqualified from the ranking results regardless of high scores in other dimensions. An optional "Flag" status may indicate borderline cases requiring manual candidate review before final rejection.

Can the dimension weights be customized in the framework?

The weights are hardcoded in lines 34-38 of .claude/skills/job-application-assistant/04-job-evaluation.md. While the framework does not expose runtime configuration for these values, users can modify the percentages directly in this skill definition file. The tools/check_framework_version.py utility ensures version tracking when such modifications occur.

What are the pre-scoring gates mentioned in the evaluation process?

Before calculating the five-dimensional scores, the framework applies an Eligibility gate and a Language gate. These preliminary checks validate fundamental requirements such as work authorization and language proficiency. Only job postings passing both gates proceed to the weighted dimensional scoring and final categorization.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}