Tanova AI Skills Taxonomy for Technical Hiring & Recruitment
Free, open-source skills framework for candidate evaluation. 139 technical skills with AI-powered matching for smarter hiring. Download JSON, use our API, or try live—no signup required.
What is a Skills Taxonomy?
A skills taxonomy is a structured framework that organizes and categorizes skills in a systematic way—think of it as a "library classification system" for professional capabilities. Just as libraries organize books by subject, author, and genre to help you find what you need, a skills taxonomy organizes technical and professional skills into hierarchical categories with standardized names and relationships.
Unlike simple keyword lists, a proper skills taxonomy includes:
- Canonical names: One official name per skill (e.g., "JavaScript" is the standard)
- Aliases: All variations mapped to the canonical name ("JS", "ECMAScript", "ES6" → JavaScript)
- Relationships: Parent/child connections (Next.js requires React, React requires JavaScript)
- Proficiency levels: Clear definitions of beginner, intermediate, advanced, and expert markers
- Context: Industry demand, typical roles, and prerequisites for each skill
For modern hiring and recruitment, skills taxonomies solve a critical problem: recruiters and AI systems need a common language to evaluate candidates. Without standardization, "React developer" might mean completely different things to different people, and automated screening systems can't make intelligent inferences about candidate qualifications.
Why Tech Recruiters Need Specialized Taxonomies
Not all skills taxonomies are created equal. Generic taxonomies like Lightcast's 34,000+ skills or ESCO's 13,000+ competencies try to cover every possible skill across every industry—from "Microsoft Excel" to "Forklift Operation" to "React Development."
The problem? More isn't better when it creates noise instead of signal.
Tanova focuses on 139 carefully curated technical skills for modern software development, with unprecedented depth:
- 556 distinct skill-level combinations (139 skills × 4 proficiency levels)
- 4 proficiency tiers with detailed behavioral markers for each level
- AI-powered transferability scores (0.0-1.0) showing skill relationships
- Tech-stack context: Prerequisites, child skills, related skills, typical roles
When screening candidates for a React role, you don't need to wade through 34,000 skills where "PowerPoint" has equal weight to "TypeScript." You need a taxonomy built specifically for technical recruitment, where every skill matters and relationships are precisely mapped for accurate candidate evaluation.
The Problem with Resume Keyword Matching
Traditional ATS systems treat skills as raw text strings, causing recruiters to miss qualified candidates:
- "React" vs "React.js" vs "ReactJS" - treated as different skills
- No skill inference - candidates must list every possible variation
- Transferability ignored - "Next.js expert" doesn't imply "React knowledge"
- No context awareness - "Java" could mean programming or JavaScript
Skills Taxonomy Solution for Recruiters
A structured, machine-readable taxonomy that enables intelligent candidate matching and skill inference—helping you find qualified candidates who use different terminology.
Skill Inference in Action
Example: Job requires "React" and "TypeScript"
Traditional ATS
Candidate CV lists: "Next.js, JavaScript, Vue.js"
REJECTED - Missing "React" and "TypeScript" keywords
Skills Taxonomy
Inference: "Next.js" → implies React, Node.js, JavaScript
"JavaScript" + modern frameworks → likely TypeScript capable
MATCHED - Skills inferred through transferability
How Technical Recruiters Use Skills Taxonomy
Skills taxonomies transform how recruiters evaluate technical candidates. Instead of rejecting resumes because of minor keyword differences, you can identify truly qualified candidates based on transferable skills and proven capabilities.
Automatically match candidates who use "React.js" when job posts say "React." Reduce false negatives by 40%+.
When candidates list "Next.js," automatically infer React, Node.js, and JavaScript proficiency based on skill relationships.
Compare candidates objectively using standardized skill names and proficiency levels—reducing bias in screening.
Real recruiter impact: Reduce time-to-hire by 30% by finding qualified candidates faster. Improve candidate quality by matching on actual skills, not just keywords.
AI-Powered Transferability Scoring for Candidate Matching
What sets Tanova apart from traditional taxonomies is our AI-powered transferability scoring system. Each skill has precisely calculated similarity scores (0.0 to 1.0) showing how closely it relates to other skills.
This enables automatic skill inference that goes far beyond simple keyword matching. When a candidate lists "Next.js" on their resume, our system automatically infers related skills with specific confidence levels.
{
"id": "javascript",
"canonical_name": "JavaScript",
"aliases": ["JS", "ECMAScript", "ES6"],
"transferability": {
"typescript": 0.95,
"python": 0.6,
"java": 0.5
},
"child_skills": [
"react", "vue", "nodejs"
],
"prerequisites": ["html", "css"]
}JavaScript → TypeScript (0.95)
Nearly identical skills, minor syntax differences
JavaScript → Python (0.6)
Similar concepts, different ecosystems
JavaScript → Java (0.5)
Shared name only, different paradigms
Real-world impact: A candidate with "Next.js" experience automatically receives high confidence scores for React (implied by Next.js), Node.js (Next.js backend), and TypeScript (modern Next.js default)—even if they didn't explicitly list these skills.
4 Proficiency Levels with Behavioral Markers
Each of our 139 skills includes detailed proficiency markers for four experience levels. This goes far beyond generic "beginner/advanced" labels—we define specific behavioral indicators that distinguish each level.
Beginner (0-1 years)
Learning fundamentals and building basic projects
• Written simple scripts
• Followed tutorials
• Basic DOM manipulation
• Used variables, loops, functions
Intermediate (1-3 years)
Building complete applications independently
• Built complete web applications
• Worked with async/await and promises
• Used ES6+ features
• Debugged complex issues
Advanced (3-5 years)
Architecting systems and mentoring others
• Architected scalable applications
• Deep understanding of closures, prototypes
• Performance optimization
• Mentored junior developers
Expert (5+ years)
Contributing to the ecosystem and setting standards
• Contributed to JavaScript ecosystem
• Deep knowledge of V8 engine internals
• Authored popular libraries
• Conference speaker on JS topics
Why this matters: Generic taxonomies might list "JavaScript: Beginner/Advanced" with no clear definition. Our behavioral markers create consistent, objective criteria for assessing skill levels—reducing bias and improving hiring accuracy.
How Tanova Compares to Other Taxonomies
When choosing a skills taxonomy, depth matters more than breadth for technical hiring. Here's how we compare to major alternatives:
| Feature | Tanova | Lightcast | ESCO | O*NET |
|---|---|---|---|---|
| Total Skills | 139 | 34,000+ | 13,500+ | 1,000+ |
| Focus Area | Tech/Software | All Industries | All Industries | All Industries |
| Proficiency Levels | 4 (with markers) | Basic only | None | 5-7 scales |
| Transferability Scores | ✓ (0.0-1.0) | ✗ | ✗ | ✗ |
| AI-Powered Inference | ✓ | ✗ | ✗ | ✗ |
| Parent/Child Relations | ✓ (detailed) | ~ (basic) | ~ (basic) | ~ (basic) |
| Aliases/Variations | 430+ | Included | Limited | Limited |
| License | CC BY 4.0 | Paid API | CC BY 4.0 | Public Domain |
| Last Updated | Dec 2025 | Ongoing | 2023 | Ongoing |
| Best For | Tech hiring, AI inference | Broad market data | EU compliance | US labor research |
The bottom line: If you're building hiring tools for software developers, you need precision, not just coverage. Tanova's tech-focused approach with AI-powered transferability scoring delivers smarter matches than generic taxonomies with thousands of irrelevant skills.
What's Included
Documentation
Complete taxonomy structure, usage examples, and integration guide
JSON Data
139 skills with 430+ aliases in machine-readable JSON format
Usage Examples
Sample code for normalization, inference, and matching
Contributing Guide
How to add skills, report issues, and submit improvements
License
CC BY 4.0 - Free to share, adapt, and use commercially with attribution
Changelog
Version history and taxonomy evolution
For Companies: See This Taxonomy in Action
We use this skills taxonomy to power intelligent candidate screening. See what your hiring team would see when evaluating applicants.
Skill Inference
Next.js on CV → Automatically infers React, Node.js, JavaScript proficiency
Hidden Gem Detection
Flags candidates with high potential but non-traditional backgrounds
Try It Yourself: Apply to a Real Job
Experience our skills taxonomy in action. Apply to Pedersen Development's Fractional CTO talent pool and see how the system evaluates your technical skills using intelligent skill inference.
Lists "React.js"? We recognize it as React
Next.js experience → React knowledge inferred
Get your 7D evaluation report immediately
Free to try • No commitment • Instant results
Book a 20-Minute Demo
See how our skills taxonomy powers intelligent candidate screening. Perfect for ATS developers, recruiters, and hiring teams.
Frequently Asked Questions
›What's the difference between a skills taxonomy and a skills ontology?
While often used interchangeably, these terms have distinct meanings:
- Skills Taxonomy: A hierarchical classification system that organizes skills into categories and subcategories (like a family tree). Focuses on structure and organization.
- Skills Ontology: A more complex system that defines skills, their properties, and the relationships between them (like a knowledge graph). Focuses on relationships and meaning.
Tanova combines both approaches: we provide taxonomic structure (categories, hierarchies) and ontological depth (transferability scores, parent/child relationships, proficiency definitions).
›Why only 139 skills instead of thousands like other taxonomies?
Quality and depth over quantity. Here's why fewer, deeper skills work better for technical hiring:
- Signal vs noise: Generic taxonomies with 34,000+ skills dilute focus. Do you really need "Microsoft PowerPoint" and "React" in the same system?
- Maintenance burden: More skills = more outdated information. We can keep 139 skills current and accurate.
- Deep proficiency data: Each of our 139 skills has 4 detailed proficiency levels (556 combinations), transferability scores, and relationship mapping—depth that's impossible to maintain across thousands of skills.
- Tech-focused: We cover modern software development comprehensively. For other domains, generic taxonomies are better.
Think of it like Michelin restaurants vs. Yelp. Fewer, carefully curated entries with deep analysis beats massive, shallow coverage.
›How often is the taxonomy updated?
We update the taxonomy on a quarterly basis (every 3 months) to reflect:
- New frameworks and technologies gaining industry adoption
- Deprecated or declining technologies (removed or marked as legacy)
- Updated proficiency markers based on evolving best practices
- Community suggestions from GitHub issues and pull requests
- Refined transferability scores based on real-world matching data
Current version: v1.4.11 (last updated December 2025). All updates are documented in our CHANGELOG with version history.
›Can I integrate this with my ATS or hiring platform?
Yes, absolutely! The taxonomy is designed for easy integration:
Three integration options:
- Direct JSON import: Download taxonomy.json and import into your database. Parse the JSON structure to access skills, aliases, relationships, and proficiency data.
- API access: Enterprise customers can use our REST API for real-time skill matching and inference (contact us for API documentation and pricing).
- Custom implementation: Fork the GitHub repo, adapt to your needs, and maintain your own version (CC BY 4.0 allows this).
The JSON structure is well-documented with examples in our GitHub README. Most developers can integrate in under a day.
›What makes your transferability scoring unique?
Our transferability scores (0.0-1.0) are AI-calculated based on real-world skill relationships, not manually guessed. This makes them:
- Objective: Derived from analysis of job postings, project requirements, and tech stack dependencies—not human opinion
- Precise: Scored from 0.0 (no relationship) to 1.0 (near-identical), allowing nuanced matching
- Directional: JavaScript → TypeScript (0.95) is different from TypeScript → JavaScript (0.85)—one direction is easier than the other
- Context-aware: Takes into account prerequisite skills, typical learning paths, and ecosystem overlap
Example: TypeScript = 0.95 transferability from JavaScript because TS is a superset with similar syntax. But Java = only 0.5 despite the similar name—different ecosystems, paradigms, and use cases.
This level of granularity enables AI systems to make intelligent inferences that simple "related skills" lists cannot.
›Where can I find affordable AI solutions for skills taxonomy?
Tanova offers both free and affordable options:
- 100% Free: Download our open-source taxonomy (139 skills, 430+ aliases) from GitHub. Use it in your own tools with no cost or attribution required beyond CC BY 4.0.
- €2/evaluation: Use our AI-powered screening that applies the taxonomy + 7D Framework for deep skill analysis. No monthly fees, pay per use.
- Custom pricing: Enterprise API access for high-volume hiring (contact for quote).
Most affordable option: Download the free taxonomy and implement your own matching logic.
›What is an AI skills taxonomy directory?
An AI skills taxonomy directory is a structured database of technical skills that includes:
- Canonical skill names: Standardized names for each skill (e.g., "React")
- Aliases: Variations like "React.js", "ReactJS", "React JS" all mapped to "React"
- Relationships: Which skills imply others (e.g., Next.js → React, Node.js)
- Categories: Groupings by technology type (frontend, backend, cloud, etc.)
- AI integration: Machine-readable format for automated skill matching
This enables hiring systems to match candidates even when they use different terminology than job descriptions.
›How do I access the AI skills taxonomy directory?
Three ways to access:
- GitHub download: Download taxonomy.json - Free, no signup, instant access
- Live demo: Try our free CV checker to see the taxonomy in action
- API access: Enterprise users can integrate our API for real-time skill matching (book a demo)
›What's the difference between this and other skills taxonomies?
Key differentiators:
- AI-optimized: Designed specifically for LLM integration and automated matching
- Transferability mapping: Shows which skills imply others (rare in public taxonomies)
- Tech-focused: Deep coverage of modern tech stack (vs. generic skills like "communication")
- Battle-tested: Used in production for 1000+ candidate evaluations
- Truly open: CC BY 4.0 license, not just "view-only" like many proprietary systems
›Can I use this taxonomy for commercial purposes?
Yes, absolutely! The taxonomy is licensed under CC BY 4.0, which means you can:
- Use it in commercial products (ATS, hiring tools, etc.)
- Modify and adapt it to your needs
- Share and redistribute it
- Build paid services on top of it
Only requirement: Give appropriate credit to Tanova in your documentation or about page.
Who Uses This Taxonomy?
Technical Recruiters
Screen candidates smarter with AI-powered skill matching. Stop missing qualified candidates due to keyword mismatches.
See how Tanova helps recruitersATS Developers
Build intelligent skill matching that goes beyond keyword search. Reduce false negatives and improve candidate quality.
Integration guideHR Tech Companies
Standardize skill representation across recruitment products. Enable better analytics and benchmarking with normalized data.
API documentationSee Skills Taxonomy in Action for Recruiting
Upload a candidate CV and job description to see how our skills taxonomy enables intelligent matching beyond keyword search. Perfect for recruiters and hiring teams.