Contextualizing Airbnb by Building a Knowledge Graph (Airbnb Engineering)

Overview

| Field | Value | |-------|-------| | Title | Contextualizing Airbnb by Building Knowledge Graph | | Source | https://medium.com/airbnb-engineering/contextualizing-airbnb-engineering-contextualizing-airbnb-by-building-knowledge-graph-b7077e268d5a | | Type | Industry engineering blog | | Category | RAG |

This entry corresponds to an Airbnb Engineering blog post on building a knowledge graph to contextualize Airbnb's products. The original article details how structured knowledge over Airbnb's domain—listings, locations, categories, and related entities—supports search, personalization, and retrieval-augmented generation (RAG) workloads.

Key Points

• Airbnb constructs a knowledge graph to provide contextual, semantic understanding beyond keyword or embedding-only retrieval.
• The knowledge graph serves as a structured external knowledge source that can be combined with LLM-based generation, aligning with the GraphRAG paradigm.
• The work belongs to the broader trend of integrating knowledge graphs with retrieval-augmented generation for industrial-scale search and recommendation systems.

Context within Search / Rec / RAG Systems

From a systems perspective, the work addresses how to redistribute responsibilities among retrieval, ranking, generation, and tool calling in the LLM era. Relevant engineering considerations discussed in the broader entry include:

• Data: PII handling, partitioned indexes, and versioned, rollback-capable embeddings.
• Latency: cascade retrieval with early stopping, caching popular queries, asynchronous reranking.
• Quality: interleaving experiments, human audits, citation verification to ensure offline gains translate to online satisfaction.
• Safety: source whitelisting, adversarial detection, and output filtering for open retrieval.
• Cost: routing to smaller models, distillation, and hybrid sparse+dense retrieval.

Related Entries

• A Survey of Graph Retrieval-Augmented Generation for Customized Large Language Models (arXiv:2501.13958)
• A Survey on Retrieval-Augmented Text Generation for Large Language Models (arXiv:2404.10981)
• Agentic Retrieval-Augmented Generation: A Survey on Agentic RAG (arXiv:2501.09136)
• RAFT: Adapting Language Model to Domain Specific RAG (Jul 2024)
• RAG vs. GraphRAG: A Systematic Evaluation and Key Insights (arXiv:2502.11371)
• Algolia's Knowledge Graphs and Ontologies — Adding Knowledge to Keyword Search

Glossary

| Term | Meaning | |------|---------| | RAG | Retrieval-Augmented Generation | | GraphRAG | RAG grounded in a knowledge graph rather than plain text corpora | | nDCG | Normalized Discounted Cumulative Gain, a ranking quality metric | | Agentic Search | Modeling search as sequential decision-making and tool invocation |

> Note: Detailed architecture descriptions and quantitative results are available in the original Airbnb Engineering blog post linked above; this page is an index entry with editorial context.