Most companies talk about vector search.
TL;DR · AI Summary
Most companies discuss vector search, but few share what it actually takes to scale to 100M+ embeddings in production.
Key Takeaways
- Booking.com used OpenSearch for initial keyword matching and later migrated to W
- Weaviate supports complex filtering, high concurrency, and write-time reads, mak
- Booking.com evaluated Weaviate's performance and cost efficiency through tests a
Outline
Jump quickly between sections.
Most companies discuss vector search, but few share what it actually takes to scale to 100M+ embeddings in production.
Basak Eskili discusses Booking.com's transition from keyword matching to large-scale production systems.
Using OpenSearch for keyword matching, handling large numbers of embeddings.
Migrating to Weaviate to handle complex filtering, high concurrency, and production-scale demands.
Weaviate's application in Booking.com's partner-to-guest message agent.
Weaviate provides relevant response templates, API fetching property and booking context, intelligent responses or human intervention (human-in-the-loop design!).
Evaluating offline datasets, LLM as a judge, A/B testing, and live partner feedback.
Booking.com exploring personalized travel agents with memory systems capturing user preferences, session context, and long-term personalization.
Mindmap
See how the topics connect at a glance.
查看大纲文本（无障碍 / 无 JS 友好）
- Weaviate 在 Booking.com 中的应用
- 初始阶段
- 使用 OpenSearch 进行关键词匹配
- 大规模扩展
- 迁移到 Weaviate 处理复杂过滤和高并发
- Weaviate 的应用
- 合作伙伴到客消息代理
- 未来展望
- 个性化旅行代理
Highlights
Key sentences worth saving and sharing.
Booking.com used OpenSearch for initial keyword matching and later migrated to Weaviate to handle large-scale needs.
Weaviate supports complex filtering, high concurrency, and write-time reads, making it suitable for large-scale production systems.
Booking.com evaluated Weaviate's performance and cost efficiency through tests and looked ahead to personalized travel agents with memory systems capturing user preferences, session context, and long-
Few share what it actually takes to scale to 100M+ embeddings in production.
Başak Eskili from @bookingcom joined the Weaviate Podcast to break down their AI journey, and it's packed with insights about what building production systems at massive scale actually looks like." / X
Most companies talk about vector search. Few share what it actually takes to scale to 100M+ embeddings in production. Başak Eskili from
joined the Weaviate Podcast to break down their AI journey, and it's packed with insights about what building production systems at massive scale actually looks like. 𝗧𝗵𝗲 𝗘𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻: • Started with keyword matching → semantic retrieval with 𝗢𝗽𝗲𝗻𝗦𝗲𝗮𝗿𝗰𝗵 on AWS • Scaled to hundreds of millions of embeddings with strict latency requirements • Migrated to 𝗪𝗲𝗮𝘃𝗶𝗮𝘁𝗲 to handle complex filtering, rising concurrency, and production-scale demands 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗚𝗲𝗻𝗔𝗜 𝗶𝗻 𝗔𝗰𝘁𝗶𝗼𝗻: Their partner-to-guest messaging agent is a real-world example of 𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜: • 𝗪𝗲𝗮𝘃𝗶𝗮𝘁𝗲 retrieves relevant response templates • 𝗔𝗣𝗜𝘀 fetch property and booking context • The agent suggests templates, crafts grounded replies, or defers to humans (human-in-the-loop design!) • Evaluation spans offline datasets, LLM-as-a-judge, A/B testing, and live partner feedback
and Başak talks about how 𝗕𝗼𝗼𝗸𝗶𝗻𝗴.𝗰𝗼𝗺 tested with 100 million embeddings, filtered vector search, multi-threaded concurrency, reads during writes, and cost-efficient infrastructure provisioning to evaluate Weaviate, as well as a look ahead at personalized travel agents with memory systems that capture user preferences, session context, and long-term personalization! Watch the full podcast here: youtube.com/watch?v=O9edM9