5
0Welcome to the Innovator Coffee, a podcast that bridges the gap between people and the world of AI and innovation. Follow us to explore the top AI products, ecosystem insights, and the emerging trends.
This episode explores how vector databases are evolving from “RAG serving infrastructure” into a broader vector lake / multimodal lake layer that can support training, multimodal data management, and agentic context engineering.
The guest explains why context engineering increases the need for better retrieval + processing pipelines, and why the market is likely to consolidate to a few survivors. Practical advice is given for builders: prioritize data + evaluation discipline, keep stacks simple early, and treat security (RBAC + encryption + key management) as essential even for small teams.
The conversation closes with a candid look at open source as GTM and the “hard, unsexy breakthrough” of stability that takes years of compounding improvements.
James (also goes by “Xiaofan Luan” in the recording) is the VP of Technology, a leading vector database company Zilliz. He and his team open-sourced Milvus, described as one of the most popular open-source vector database projects on GitHub, and he has been building Milvus for about six years. Before that, he contributed to multiple open-source database projects.
00:02 Opening & Guest Introduction
00:45 From Vector DB to Vector Lake / Multimodal Lake
03:06 Context Engineering, Memory Layers & Agents
05:41 AI Infra Outlook & Vector DB Market Consolidation
09:09 Improving Accuracy: Data, Evaluation & Context
12:47 Vector DB + MongoDB/Postgres & Startup Tradeoffs
15:37 Security in Vector Databases: RBAC, Encryption & Cloud
21:53 Open Source GTM, Stability as Moat & Closing Advice
Special thanks to Hannah Wang who did the wonderful job to assist to complete this podcast. / hannah-wang-9302421b3 and Vicky who helps editing the podcast vickylan0004@gmail.com
Hosts:Tom Kong*Stanford EE alumni,*Founder@ Stanford AGI Adventist Community (10K+ members so far from top VC, Engineers, startups from Silicon Valley )*AI Lecturer, a serial entrepreneur in media and data. Advisor @ techtimes.com*AI deployment for 8 years, with NLP and recent LLMs (RAG, Agent, Diffusion)Wickey Wang*IT Security Compliance Leader & University Faculty*Growth fund VC advisor, VC fellow and Angel Investor with cybersecurity and AI focus*GAI Security book co-author
Questions, Suggestions, Feedback and Comments? You can find us in LinkedIn:www.linkedin.com
