{
  "id": "skill-area:llm-evaluation",
  "_kind": "SkillArea",
  "_file": "domain/skill-areas/skill-areas-expanded.yaml",
  "_cluster": "domain",
  "attributes": {
    "displayName": "LLM Evaluation",
    "description": "Techniques for evaluating large language model quality, including\nautomated benchmarks, human evaluation, and domain-specific metrics.\nCovers BLEU/ROUGE, LLM-as-judge, Elo rating, and eval harness design.\n",
    "expertiseLevels": [
      "intermediate",
      "expert"
    ]
  },
  "outgoingEdges": [
    {
      "from": "skill-area:llm-evaluation",
      "to": "specialization:ai-agents-conversational",
      "kind": "applies_to",
      "attributes": {
        "confidence": "primary"
      }
    },
    {
      "from": "skill-area:llm-evaluation",
      "to": "domain:ml-ai",
      "kind": "applies_to",
      "attributes": {
        "confidence": "secondary"
      }
    },
    {
      "from": "skill-area:llm-evaluation",
      "to": "skill-area:eval-driven-development",
      "kind": "prerequisite_for_learning",
      "attributes": {
        "strength": "recommended"
      }
    },
    {
      "from": "skill-area:llm-evaluation",
      "to": "skill-area:bias-fairness-analysis",
      "kind": "prerequisite_for_learning",
      "attributes": {
        "strength": "helpful"
      }
    }
  ],
  "incomingEdges": [
    {
      "from": "skill-area:synthetic-data-generation",
      "to": "skill-area:llm-evaluation",
      "kind": "prerequisite_for_learning",
      "attributes": {
        "strength": "helpful"
      }
    },
    {
      "from": "skill-area:AI-agent-evaluation",
      "to": "skill-area:llm-evaluation",
      "kind": "prerequisite_for_learning",
      "attributes": {
        "strength": "recommended"
      }
    }
  ]
}