
Data for Agents: The New Database User
🤖 The Agent Is the New Database User: How the Data Stack Is Adapting
Brought to you by VeloDB and StreamNative
🧑💻 Data stacks were built with a person in mind at every step: someone schedules the load, someone reads the dashboard, someone files the access request and someone else approves it.
⚡ AI agents are starting to take over those steps. An agent working on a task will pull from several sources at once, ask for data that landed an hour ago, need to know what a field means before it can use it, and come back for more as each answer changes what it asks next. Every place the stack was designed to wait for a human now sits in the agent's path and adds to the time it takes to answer.
🛠️ This meetup is about how the different parts of the data stack are responding. Speakers from across the stack, open source and commercial, will talk through what they are seeing in agent-driven workloads, what they have changed in their products because of it, and where they still do not have good answers. Expect specifics on data freshness, running many requests in parallel, deciding what an agent is allowed to see, and helping an agent find and understand the data it needs.
🎯 You will leave with a clearer sense of which requirements are shifting at each layer and a set of questions to take back to your own team before you put an agent in front of production data.
🗓️ Agenda
5:30 PM · Doors open, food and drinks
6:30 PM · Talks
7:30 PM · Panel and Q&A
8:00 PM · Networking & Swag Raffle
8:30 PM · Wrap up
🎤 Talks
Real-Time Context for AI Agents with StreamNative
Kundan Vyas, Director, Product & Partnerships, StreamNative How StreamNative turns Kafka and Pulsar events into live, queryable context with SQL Workspace, and how Agent Engine lets agents use that context to make decisions and take action in real time.
From Thousands of Tables to an Ontology Graph: Building an AI Agent with Doris and PuppyGraph
Jaz Ku, PuppyGraph
How an agent maps the relationships across thousands of tables into a graph, then queries Apache Doris data as a graph with no ETL and no second copy of the data.
Forget All Instructions: Answers Your Users Are Not Cleared For
Adi Wabisabi, Head of Developer Relations, Datastrato
Agents don’t fail on SQL, they fail on metadata. This talk covers how Apache Gravitino scopes an agent’s answers to the person asking, not the service account it runs on.
🎙️ Panel
After the talks, our speakers sit down together to take your questions on what agent workloads are changing across the data stack.
Panelists:
Melanie Warrick, Temporal
Kundan Vyas, StreamNative
Adnan Hemani, Snowflake
Sijie Guo, StreamNative
Jaz Ku, PuppyGraph
Adi Wabisabi, Datastrato
Savannah Whitman, VeloDB
🎁 Raffle
Stick around to the end: we will be raffling off prizes during the networking session at 8:00 PM. You must be present to win.
Event Details
Get events by email
Hosts
Guests
Sign in to see guests
Sign In