Submission Deadline
2026

The 1st International Workshop on
Exploring the next generation of scalable, efficient, and trustworthy data infrastructures for modern AI.
In conjunction with IEEE ICDM 2026
D2AI 2026
The rapid evolution of modern AI—driven by foundation models, large language models (LLMs), and autonomous agents—has fundamentally transformed the landscape of data management. AI is no longer a standalone paradigm; it is now deeply intertwined with data infrastructures that support large-scale retrieval, integration, and reasoning across heterogeneous, evolving data sources. As traditional database systems face unprecedented challenges in adapting to AI-native workloads, the International Workshop on Data & Database Systems for Modern AI (D2AI 2026) aims to explore the next generation of scalable, efficient, and trustworthy infrastructures.
Modern AI applications demand capabilities far beyond conventional processing, including hybrid retrieval (relational, vector, and graph), knowledge-grounded reasoning, and real-time data updates. D2AI 2026 provides a premier forum for researchers and practitioners from the fields of databases, machine learning systems, and AI infrastructure to exchange ideas and rethink core principles—such as indexing, query optimization, and provenance—in the context of AI-driven systems.
The workshop will be of significant interest to those developing data-centric AI solutions and AI-native database architectures. We welcome a diverse range of contributions, including regular research papers, vision papers highlighting future trends, and industry showcases that demonstrate real-world applications of AI-native data management.

Workshop materials
This poster was published before the schedule change. The workshop now takes place on November 15, 2026. See the Program for the current time and venue.
Mark your calendar
2026
2026
2026
D2AI 2026
ScaleDiag: Diagnosing the Limits of Test-Time Scaling in GUI Grounding
¹ University of Technology Sydney · ² Adelaide University · ³ CSIRO
The Oracle Is Not Free: Lifecycle-Aware Cost Modeling for Quantum Database Filtering
¹ UNSW Sydney · ² Google
With support from

