Skills Normalization

How intelligent skills matching works

All parsed resumes include normalized skills matched against a 32,000+ skill taxonomy.

How It Works

1. Skill Extraction

The AI model extracts skills from:

Explicit skills sections
Job descriptions
Project descriptions
Certifications

2. Normalization

Each extracted skill is matched to our standard taxonomy:

Exact matches: "JavaScript" → "JavaScript"
Aliases: "React JS", "ReactJS" → "React"
Semantic matching: "Node development" → "Node.js"

3. Proficiency Levels

Skills include proficiency based on context:

Basic: "familiar with", "exposure to"
Intermediate: "working knowledge", "proficient"
Advanced: "strong experience", "skilled"
Expert: "expert", "mastery", "architected"

Proficiency considers:

Frequency of mentions
Recency of experience
Context words (senior, lead, expert)

Output Format

{
  "skills": [
    {
      "skill": "JavaScript programming",
      "normalized": "JavaScript",
      "proficiency": "expert",
      "category": "Programming Languages",
      "skill_type": "hard",
      "confidence": 0.95
    }
  ]
}

Fields

skill: Original text from resume
normalized: Standardized skill name
proficiency: basic | intermediate | advanced | expert
category: Skill category (e.g., "Programming Languages")
skill_type: hard | soft
confidence: Match confidence (0-1)

Skill Categories

Common categories include:

Programming Languages
Web Frameworks
Databases
Cloud Platforms
DevOps Tools
Project Management
Communication
Leadership

Benefits

Consistency

Different candidates writing "React", "ReactJS", "React.js" all normalize to "React" for easy comparison.

Searchability

Search your candidate database by normalized skills, not fragmented variants.

Analytics

Track skill trends and demands using consistent taxonomy.

Accuracy

Match rate: >95% of skills successfully normalized
Confidence threshold: Only matches >0.75 confidence included
Fallback: Unmatched skills included as-is with flag

Using Normalized Data

Compare Candidates

// Both candidates have same normalized skill
candidate1.skills.find(s => s.normalized === "React")
candidate2.skills.find(s => s.normalized === "React")

Filter by Proficiency

// Find expert-level candidates
const experts = candidates.filter(c =>
  c.skills.some(s =>
    s.normalized === "React" &&
    s.proficiency === "expert"
  )
)

Aggregate Skills

// Count skill frequency across all resumes
const skillCounts = resumes
  .flatMap(r => r.skills)
  .reduce((acc, s) => {
    acc[s.normalized] = (acc[s.normalized] || 0) + 1
    return acc
  }, {})

Coming Soon

Custom skills taxonomy for Professional tier subscribers (define your own mappings and categories).