Mike Liu
Mike is one of the earliest engineers on the OceanBase team. He has worked across the core database engine, from early architecture design to leading the high availability and consensus layer. He now heads the seekdb team, building an AI-native database for agentic workflows and modern AI applications.
Read more articles from this author
Why Your Vector Database Benchmark Is Wrong for AI Agents
Under streaming AI workloads, vector databases see high P99 jitter (1.1×–10.3×) under concurrency. seekdb v1.3.0’s fixed delta+snapshot HNSW avoids this, delivering 22× QPS and 19× P99 gains over prior version.
5 Minutes to a Working AI Prototype: A Developer's Guide to seekdb D0
A hands-on guide to seekdb D0, the free trial entry point for OceanBase's AI-native hybrid search database. Spin up a MySQL-compatible instance with vector, full-text, and data branching in one curl command — no signup, no credit card.
FORK TABLE: Get Your Database Ready for Agents
AI agents write to databases at high frequency with no accountability. FORK TABLE in seekdb gives every agent its own sandbox — branch instantly, experiment safely, discard or promote results.
Stop Stitching, Start Building: Get Started with OceanBase seekdb
In this post, we will skip the architecture talk and go straight to code. You go from a blank editor to a fully operational, ACID compliant hybrid search engine running entirely in user space.
From Complex to Simple: How We Built seekdb for the AI Era
AI era doesn't need another heavy, complex enterprise database. It needs agility. It needs flexibility. We went back to the drawing board to understand what an AI application actually needs from a database. Our answer is OceanBase seekdb