How Airbnb Rebuilt Its Search for Instant Bookings

In this episode of The CTO Podcast, Lucas and Luna dive into how Airbnb rewrote its search backend to handle instant bookings at global scale. They explore the specific engineering challenge: reducing search latency from 2.3 seconds to under 200 milliseconds while adding real-time pricing and availability filters. The conversation covers the architectural shift from a monolithic Rails app to a bespoke C++ search service using Apache Lucene, and the trade-offs between index freshness and query speed. Lucas explains why Airbnb chose not to use Elasticsearch, how they handled the 'cold start' problem for new listings, and the surprising role of machine learning in reranking results. Luna pushes back on the cost of custom infrastructure versus managed services. Perfect for anyone building real-time search or scaling a marketplace platform. #Airbnb #SearchArchitecture #RealTimeSearch #ApacheLucene #CPlusPlus #Microservices #LatencyOptimization #MarketplaceTech #EngineeringTradeoffs #MachineLearning #IndexFreshness #SearchReranking #BusinessAndTechnology #CTOPodcast #FexingoBusiness #BusinessPodcast #TechnicalLeadership #EngineeringOrg Keep every episode free: buymeacoffee.com/fexingo