Skein: An Embedded Knowledge Database for the AI Era

Keynote
Shanghai
2:10 p.m. - 2:50 p.m.
Venue C(Integrated Building, Room 421 Meeting Room)
Talk Poster
  • Weizhen Wang Co-founder & CTO, Nowledge Labs

    Currently dedicated to exploring and building cutting-edge Agent Memory and Harness systems. Former core engineer of the PingCAP TiDB kernel optimizer and the project's No.1 contributor. With many years in foundational software, he has deep expertise and hands-on experience in distributed system architecture, database kernels, low-level storage, and infrastructure.

    wangweizhen

Abstract

Starting from why Skein was created, this talk shares thoughts on what databases should look like in the AI era, and the new needs, challenges and possibilities that appear when databases start serving “knowledge” rather than just “data”.

Details

AI applications are redefining the value of data: data is no longer merely stored and queried — it needs to be understood, connected, and gradually shaped into knowledge through continuous interaction. Skein is an embedded knowledge database we built around this shift, aiming to provide a more natural data foundation for AI applications, agents, and personal knowledge management. Starting from the background of Skein’s creation, this talk introduces our thinking about the shape of databases in the AI era, and the explorations, trade-offs and practical lessons from its design and implementation. It is not a complete technical teardown; rather, through Skein, it discusses the new needs, challenges and possibilities that emerge when databases begin to serve “knowledge” instead of just “data”.