Agentic Data Systems
The Joint Workshop on Agentic Data Systems and Data-Centric AI
Bringing together researchers and practitioners to explore system-level challenges of autonomous data workflows — from long-horizon planning and execution-grounded control loops to governance and workflow-level evaluation.
About ADS 2026
The 1st ADS & 3rd DATAI Workshop at VLDB 2026 · Boston, MA, USA
Modern data systems are growing in complexity and scale. Recent advances in large language models (LLMs) have enabled more autonomous data management capabilities. Agentic Data Systems (ADS) represent an emerging paradigm that integrates LLM-driven agents into data platforms to plan, execute, and govern end-to-end data workflows.
Unlike traditional "copilot" tools that provide assistance for individual steps, ADS aims to orchestrate complete data workflows — from data discovery and preparation to analysis and reporting — while satisfying users' requirements on correctness, cost, and compliance.
ADS 2026 brings together researchers and practitioners to explore the system-level challenges of ADS, including long-horizon planning, execution-grounded control loops, error recovery, optimization under resource budgets, and policy-aware governance.
Key Research Challenges
- Data-environment perception at scale
- Workflow abstraction & long-horizon orchestration
- Reliability under cascading errors
- Efficiency and cost-aware decision making
- Evaluation & benchmarks beyond accuracy
- Governance, safety, and accountability
Workshop History
This joint workshop combines the 1st ADS workshop with the 3rd DATAI workshop. Previous DATAI workshops at VLDB 2024 and 2025 attracted 50+ participants each year, featuring keynotes from leading researchers at SFU, Baidu, KAIST, UMN, TU Delft, and BIFOLD.
Important Dates
All deadlines are Anywhere on Earth (AoE)
Call for Papers
We welcome papers exploring the intersection of LLMs, Agents, and Data Management
Paper Types
Short papers (up to 6 pages) and long papers (up to 12 pages), excluding bibliography. Survey, vision, tutorial, demo, and late-breaking results are all welcome.
Review Process
Single-blind peer review — all author names and affiliations must be included. Each submission is reviewed by at least three Program Committee members. Papers outside the scope will be desk rejected.
Submission
Papers should be submitted via the workshop submission system. Authors must use the official VLDB Workshop style file and fill in VLDB 2026 as the conference year and the workshop name in the reference format.
Paper Formatting Requirements
All accepted papers will be included in the general VLDB 2026 Workshop Proceedings.
Authors must format their papers using the official VLDB Workshop style file. Please ensure that you fill in the correct conference year (VLDB 2026) and the workshop name (ADS 2026: The Joint Workshop on Agentic Data Systems and Data-Centric AI (The 1st ADS & 3rd DATAI)) in the VLDB Workshop Reference Format of the paper.
Download VLDB Workshop Style FileTopics of Interest
- Agentic control planes for data stacks; planner–executor–verifier; supervisor models
- Workflow representations: DAGs, declarative/typed operators, semantic operators, tool composition
- Tool/operation interfaces for DBMS, lakehouse, catalogs, pipelines, notebooks, BI, and observability
- Agent memory/logging: execution traces, intermediate artifacts, provenance, data/version-aware retrieval
- Multi-agent collaboration for data workflows (roles, communication, delegation, verification)
Tentative Program
Full-day in-person workshop — Friday, September 4, 2026
* Tentative schedule, subject to change based on VLDB 2026 official program.
Organizers
Workshop Co-Chairs
Professor, ACM/IEEE Fellow
Tsinghua University
ACM Fellow and IEEE Fellow. VLDB 2017 Early Research Contribution Award. SIGMOD 2024 Research Highlight Award. SIGMOD 2023 Best Papers. SIGMOD 2021 General Co-Chair.
Homepage
Assistant Professor
HKUST (Guangzhou) & HKUST
Research at the intersection of Data and AI, focusing on Data Agents and Data-centric AI. 50+ publications in top-tier venues (SIGMOD, VLDB, KDD, ICML, NeurIPS, ICLR). Best-of-SIGMOD 2023 recipient. KDD Cup 2026 Chair (Data Agents Track).
Homepage
Assistant Professor
Shanghai Jiao Tong University
SIGMOD 2025 Jim Gray Dissertation Honorable Mention (first from mainland China). VLDB 2023 Best Industry Paper Runner-up. CCF Outstanding Doctoral Dissertation Award.
HomepageAcknowledgment
The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.