Knowledge Graph

The Beamery Knowledge Graph (BKG) is a massive, continuously evolving semantic network that models the global talent ecosystem. With over 20 billion triples connecting people, skills, companies, and roles, it provides the intelligence foundation for all of Beamery's AI capabilities.

Overview

The Knowledge Graph represents a paradigm shift in how HR technology understands talent. Rather than treating data as isolated records, the BKG captures rich relationships and context that enable true intelligence.

What Makes It Unique

Core Value Propositions

1. Contextual Understanding - Skills mean different things in different contexts
2. Relationship Insights - Discover non-obvious connections
3. Predictive Intelligence - Anticipate talent trends and needs
4. Global + Local - Worldwide patterns with local relevance

Graph Architecture

Semantic Model

The BKG uses a sophisticated ontology based on industry standards:

Triple Structure

Each relationship in the graph follows RDF triple format:

# Person has skill
<person:123> <has_skill> <skill:python_programming> .
<person:123> <skill_proficiency> "advanced" .
<person:123> <years_experience> "5" .

# Role requires skill  
<role:software_engineer> <requires_skill> <skill:programming> .
<role:software_engineer> <typical_seniority> <seniority:mid_level> .

# Skill relationships
<skill:react> <broader_concept> <skill:javascript_frameworks> .
<skill:react> <related_to> <skill:redux> .

Named Graphs

Customer data is isolated using named graphs:

graph:beamery_global {
  # Public knowledge
  - Canonical taxonomies
  - Market intelligence  
  - Industry standards
  - Public profiles
}

graph:customer_12345 {
  # Private knowledge
  - Employee data
  - Custom taxonomies
  - Internal roles
  - Proprietary skills
}

Temporal Versioning

Track how knowledge evolves over time:

{
  "entity": "skill:angular",
  "versions": [
    {
      "valid_from": "2010-01-01",
      "valid_to": "2016-12-31",
      "name": "AngularJS",
      "version": "1.x"
    },
    {
      "valid_from": "2016-09-01",
      "valid_to": "current",
      "name": "Angular",
      "version": "2+"
    }
  ],
  "relationships": [
    {
      "type": "succeeded_by",
      "from": "skill:angularjs",
      "to": "skill:angular"
    }
  ]
}

Data Sources

Primary Sources

Data Quality Pipeline

Privacy & Compliance

Job Architectures

Customer-Specific Subgraphs

Each organization gets a tailored job architecture:

{
  "customer": "enterprise_corp",
  "job_architecture": {
    "job_families": [
      {
        "name": "Engineering",
        "roles": [
          {
            "id": "role_swe_1",
            "title": "Software Engineer I",
            "level": 1,
            "skills": ["programming", "debugging"],
            "next_role": "role_swe_2"
          },
          {
            "id": "role_swe_2",
            "title": "Software Engineer II",
            "level": 2,
            "skills": ["system_design", "mentoring"],
            "next_role": "role_senior_swe"
          }
        ]
      }
    ],
    "skill_framework": {
      "core_skills": ["collaboration", "communication"],
      "technical_skills": {
        "engineering": ["programming", "architecture"],
        "data": ["sql", "analytics"]
      }
    },
    "career_paths": [
      {
        "from": "role_swe_1",
        "to": "role_tech_lead",
        "required_skills": ["leadership", "architecture"],
        "typical_timeframe": "5-7 years"
      }
    ]
  }
}

Architecture Creation Process

1. Initial Ingestion

   • Import existing job data from HRIS
   • Map custom fields to standard schema
   • Preserve organizational structure

2. AI Enhancement

   • Enhance with market intelligence
   • Infer missing skills from descriptions
   • Suggest appropriate seniority levels

3. Collaborative Refinement

   • Review AI suggestions
   • Adjust based on organizational needs
   • Add company-specific requirements

Architecture Analytics

Gain insights from your job architecture:

Graph Capabilities

Query Patterns

Execute sophisticated queries against the Knowledge Graph:

• Entity lookup - Find specific concepts, roles, or skills
• Relationship traversal - Follow connections between entities
• Pattern matching - Discover similar structures
• Aggregate analysis - Statistical insights across the graph

Embeddings & Similarity

The graph provides pre-computed embeddings for:

• Semantic similarity - Find related concepts
• Career similarity - Identify similar career paths
• Skills clustering - Group related competencies
• Role matching - Compare job profiles

Real-time Updates

Stay current with graph evolution:

• Taxonomy updates - Quarterly skill and role updates
• Market signals - New trends and emerging skills
• Customer changes - Reflect organizational updates
• Relationship discovery - New connections identified

Best Practices

Advanced Patterns

Multi-hop Reasoning

• Find skills that lead to leadership roles
• Trace career paths backwards from target positions
• Identify common skill progressions

Market Intelligence

• Track emerging skills in your industry
• Compare skill demand year-over-year
• Identify declining competencies

Competitive Analysis

• Benchmark your job architecture
• Compare skill coverage to industry
• Assess career mobility options