The Open-Source Data Stack

​The Open-Source Data Stack for AI Agents 🤖

​Most data infrastructure was designed for human users: analysts running queries, reviewing dashboards, and investigating when results look wrong.

​AI agents use the same infrastructure differently. They issue queries continuously, at higher volume, and in patterns that were not anticipated. They cannot judge whether data is stale or incomplete, and they act on results without review. Systems that were adequate for human workloads often are not adequate for agent workloads.

​Open source is what can keep these systems flexible. An agent integrated with a single proprietary platform is expensive to migrate, and that vendor lock-in is real. Every project presented at this event can be self-hosted, forked, or replaced with an alternative. That way, teams can adopt what works, swap what doesn't, and keep control of their own stack.

​Join VeloDB (Apache Doris) for an evening on what changes at each layer of the data stack when the primary consumer is an agent. Food, drinks, and plenty of networking with other builders to be provided. 🍕

​📅 Agenda

​5:30 Doors, drinks, networking
6:00 Welcome
6:10 Lightning talks
7:10 Panel
7:55 Raffle 🎁
8:05 Networking

​📍 Wednesday, October 28 · 5:30 to 8:30pm
SV/AI, Silicon Valley AI Hub · 135 Constitution Drive, Menlo Park