A shared space for
the next data stack.

An open, vendor-neutral, practice-driven community exploring Data Agents and the next generation of data infrastructure. We connect developers, data engineers, architects, researchers, enterprise teams, and open-source communities to share what works, what fails, and what comes next.

Why we started

As AI agents become users—and even builders—of data systems, is today's infrastructure ready?

The Agentic Data Stack is a data technology stack designed and evolved for AI agents, enabling them to understand data and business context and carry out discovery, development, analysis, and operations under governed, verifiable conditions.

This shift requires more capable Data Agents and a rethink of infrastructure. We bring developers, engineers, architects, researchers, enterprise teams, and open-source communities together to share successes and failed attempts alike.

Real problems beyond product features. Production practice beyond demos. Open discussion across technical approaches, with no single prescribed answer.

Six areas of exploration

Data Agents in practice

Planning, tool use, context management, and verification in data development, analytics, and operations.

Data & business context

Help agents understand business meaning beyond SQL with reusable enterprise context.

Semantics, graphs & ontology

Connect metrics, business objects, relationships, and rules to move from analysis toward decisions and action.

Agent-ready infrastructure

How databases, catalogs, lakehouses, and pipelines evolve for autonomous exploration.

Reliability in production

Evaluation, accuracy, permissions, security, observability, and the cost of complex execution.

People & agents

How engineers, analysts, and business teams work as agents take on more of the data workflow.

Co-initiated. Built together.

Co-initiated by DatusAI, NebulaGraph, and StarRocks. Open to contributions from companies, open-source projects, and individuals.

AI-native workflows for data teams.

An open-source engine for real-time analytics.

How we meet and organize

Meetups connect developers, technology builders, and enterprise practitioners. Hosts are identified per event; public talks, articles, and lessons are collected on this website with permission.

Help shape the community

Contribute a technical talk, a case study, an article, or your experience organizing events.