{
  "id": "skill-area:retrieval-evaluation",
  "_kind": "SkillArea",
  "_file": "domain/skill-areas/skill-areas-knowledge-fabric.yaml",
  "_cluster": "domain",
  "attributes": {
    "displayName": "Retrieval Evaluation",
    "description": "Measuring and improving RAG pipeline quality — evaluation metrics\n(faithfulness, answer relevance, context precision, context recall),\nevaluation frameworks (Ragas, DeepEval, TruLens), building golden\nevaluation datasets, A/B testing retrieval configurations, monitoring\nretrieval quality in production, and the distinction between\ncomponent-level evaluation (retriever quality) and end-to-end\nevaluation (final answer quality).\n",
    "domains": [
      "specialization:ai-agents-conversational"
    ],
    "expertiseLevels": [
      "intermediate",
      "expert"
    ]
  },
  "outgoingEdges": [
    {
      "from": "skill-area:retrieval-evaluation",
      "to": "specialization:ai-agents-conversational",
      "kind": "applies_to",
      "attributes": {
        "confidence": "primary"
      }
    },
    {
      "from": "skill-area:retrieval-evaluation",
      "to": "tool:haystack",
      "kind": "uses_tool"
    }
  ],
  "incomingEdges": [
    {
      "from": "skill-area:embedding-optimization",
      "to": "skill-area:retrieval-evaluation",
      "kind": "prerequisite_for_learning",
      "attributes": {
        "strength": "required"
      }
    },
    {
      "from": "tool:skillachi",
      "to": "skill-area:retrieval-evaluation",
      "kind": "tool_used_by",
      "attributes": {}
    }
  ]
}