Milvus Meetup Tokyo 12月3日涉谷举办:RAG 与向量搜索实践分享
Milvus 官方在东京办的线下聚会,四位讲者都来自在用 Milvus 的公司:DeNA 讲 RAG 扩展踩过的坑,还有讲从 Azure AI Search 迁移过来的真实对比。在日本做 RAG 或向量搜索的工程师值得去现场听一手经验。
Milvus 官方宣布将于 2026 年 12 月 3 日 7:00–9:30 p.m.(JST)在东京涉谷 TKP Garden City 举办 Milvus Meetup Tokyo,共四场演讲与 Q&A,报名通过 Connpass 进行。DeNA 的 Isana Kondo 将介绍 DAAQ 的 RAG 系统随规模增长后的租户设计调整,以及基于 GKE 运行 Milvus 的经验;KDDI iret 的 Kento Yamada 讨论专用向量数据库的适用场景与托管服务的差异;Daiva Institute of Research 的 Eisuke Izawa 分享 Dione Workspace 从 Azure AI Search 迁移到 Milvus 的原因;Zilliz 的 Yulan Yan 则带来大规模搜索部署案例和 Milvus 3.0 的新特性。演讲以日语进行,面向在生产环境中落地向量搜索、RAG 和 AI 智能体记忆的工程师,会后可与讲者交流各自项目中的具体问题。
对在日本市场部署 RAG 系统的团队来说,这类线下聚会的价值在于能看到真实生产案例,比如 GKE 集群运维、多租户改造和从 Azure AI Search 迁移的成本收益对比。与纯线上文档或国际会议相比,本地化案例对区域团队更具参考性。
We’re hosting 𝗠𝗶𝗹𝘃𝘂𝘀 𝗠𝗲𝗲𝘁𝘂𝗽 𝗧𝗼𝗸𝘆𝗼 in Shibuya on December 3! Join us for an evening about talks and Q&A on bringing vector search, RAG, and AI agent memory into production. • 𝗜𝘀𝗮𝗻𝗮𝗼 𝗞𝗼𝗻𝗱𝗼 (𝗗𝗲𝗡𝗔) will share how his team adapted the RAG system behind DAAQ as it grew, including changes to its tenant design and lessons from running Milvus on GKE. • 𝗞𝗲𝗻𝘁𝗼 𝗬𝗮𝗺𝗮𝗱𝗮 (𝗞𝗗𝗗𝗜 𝗶𝗿𝗲𝘁) will discuss when a dedicated vector database makes sense and what changes when you choose a managed service. • 𝗘𝗶𝘀𝘂𝗸𝗲 𝗜𝘇𝗮𝘄𝗮 (𝗗𝗮𝗶𝘄𝗮 𝗜𝗻𝘀𝘁𝗶𝘁𝘂𝘁𝗲 𝗼𝗳 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵) will explain why Dione Workspace moved from Azure AI Search to Milvus. • 𝗬𝘂𝗹𝗮𝗻 𝗬𝗮𝗻 (𝗭𝗶𝗹𝗹𝗶𝘇) will share examples from large-scale search deployments and cover what’s new in Milvus 3.0. Bring a question from your own project and stay after the talks to meet the speakers and compare notes with other engineers. 📅 Thursday, December 3, 2026, 7:00–9:30 p.m. JST 📍 TKP Garden City Shibuya, Hall 11B Talks are in Japanese. Admission milvus.connpass.com/event/412862/ av #Milvus o #VectorDatabase #OpenSource d #MilvusMeetup e your spot on Connpass: https://t.co/rIIOHn5WTk #Milvus #VectorDatabase #OpenSource #MilvusMeetup
This report is categorized under AITOP's "Products" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Milvus, Zilliz, RAG, vector databases.
Source attribution: curated from Milvus (@milvusio).