Meet the speakers! Carsten Binnig - Professor@TU Darmstadt & Visiting Researcher@SystemsResearch@Google His research on scalable data systems on modern hardware & machine learning for scalable data systems received a Google Faculty Award & multiple Best Paper & Best Demo awards

We are happy to announce that our paper "#ZeroTune: Learned Zero-Shot Cost Models for Parallelism Tuning in Stream Processing" by Pratyush Agnihotri et al. was accepted to the A* conference IEEE ICDE Conference. Download full paper here: kom.tu-darmstadt.de/publications/A…

We are happy to share that our paper "Parallel Processing of Temporal Anti-Joins in Memory" has been accepted to DASFAA'24! 🚀🚀 The paper introduces a novel approach to efficiently process temporal anti-joins, a task known for its complexity and scalability issues.

I was running my first Dagstuhl last week with David Bacon (Google), Dave Patterson (UC Berkeley), and Margo Seltzer Margo Seltzer (UBC) bringing together researchers on #CloudDatabases & #ComputerArchitecture : dagstuhl.de/24162/. The was a fun & highly interesting week.

Did you really want to attend the VLDB #SummerSchool on #DataManagement in the #Cloud but you missed the application period? Hurry, there is still a chance!

🎉 #DAMON2024 turns 20 with a fresh new look! 🌟 Check out the re-designed website and the list of accepted papers at damon-db.org

We are proud to announce that our recent paper "COSTREAM: Learned Cost Models for Operator Placement in Edge-Cloud Environments" has been accepted to ICDE'24. Roman Heinrich presents his work next week in the AI for Databases III session at ICDE. arxiv.org/abs/2403.08444

We proudly share that our paper "Rethinking Table Retrieval from Data Lakes" was accepted to the aiDM'24 workshop ACM SIGMOD/PODS. The paper proposes the idea of fine-grained structured table retrieval to take on two limitations of existing table retrievers. doi.org/10.1145/366374…

We proudly present our paper "Surprise Benchmarking: The Why, What, and How", accepted at the DBTest'24 workshop ACM SIGMOD/PODS. The work resulted from the Schloss Dagstuhl Seminar 23441 on Ensuring the Reliability and Robustness of Database Management Systems. doi.org/10.1145/366216…

I am honored that we just received the best demo award for our demo at SIGMOD/PODS 2025 2024. Congrats to my PhD student Matthias Urban Systems Group @ TU Darmstadt (Moved) for this great work.

Due to ACM SIGMOD/PODS, I totally forgot to share that Sven Langenecker of Systems Group @ TU Darmstadt (Moved) defended his PhD thesis on learned semantic type detection in tabular data lakes on June 4. Congrats, Sven! It was a pleasure working with you 🎉 🥳

It is a really great honor to receive the LOEWE top professorship for my research on building the next generation of AI-centered database systems: wissenschaft.hessen.de/presse/loewe-s… Many thanks to the State of Hesse & TU Darmstadt as well as hessian.AI and Computer Science, TU Darmstadt.

We Systems Group @ TU Darmstadt (Moved) had a great time last week in our team retreat at Darmstädter Haus in the Austrian Alps discussing the future of AI & cloud data management. Thanks a lot also to our external guests, Norman May and Ulf Brackmann from SAP.

We proudly announce that our recent paper "LLMs for Data Engineering on Enterprise Data" has won the Best Short Paper award of the TaDA Workshop VLDB 2024. The paper reveals the limitations of LLMs when faced with real-world scenarios like enterprise data. tabular-data-analysis.github.io/tada2024/paper…

I am very proud of my team who worked on this paper jointly with SAP. 🎉 The results clearly show that more work is needed to make LLMs work on enterprise data. We will continue this line of work which we hope will lead to large tabular models for enterprise data.

Happy to share that our work on WikiDBs - a large-scale corpus of relational databases - was accepted as a spotlight to the Dataset & Benchmarks track NeurIPS Conference. This is an important step for enabling foundation models on structured databases rather than text or images.

CMT for round 1 of ACM SIGMOD 2026 is now open. SIGMOD/PODS 2025 2026 has a two-step submission process: all authors must have a registered CMT account with an ORCID & completed conflict of interest declarations by the abstract submission deadline (10 Jan 2025, 11:59 PM AoE).

Abstract submission deadline of ACM SIGMOD/PODS 2026 (round 1) is in 2 days, 10 Jan, AoE. For this (1) All authors must have a registered CMT account with a valid ORCID. (2) All authors must declare their domain and their individual conflicts. Please check the CfP carefully.

Extremely proud to share that our paper "Towards Foundation Database Models" by Johannes Wehrstein (joint work with Google) has won the Best Paper Award at CIDR'25 CIDR 2025! 🎖️🚀 This work revolutionizes how we realize learned database components. vldb.org/cidrdb/papers/…

🚨 Important Update! 🚨 We've moved our social presence to Bluesky and LinkedIn! Follow us on Bluesky つだ-systems.bsky.social Connect with us on LinkedIn linkedin.com/company/tuda-s… This isn't goodbye, it's see you there! 👋 #Bluesky #LinkedIn