Launching Our Blog And Wrapping Up 2025

I’m super excited to launch our blog! We’ll use this space to share what’s happening in our lab, from research papers and systems to the day-to-day life of our team. To kick things off, let’s look back at 2025.

Award-Winning Research

We were super happy to see our work recognized by the community this year!

• SIGMOD 2025 Honorable Mention: Our paper “DPconv: Super-Polynomially Faster Join Ordering” by Mihail Stoian and Andreas Kipf received an Honorable Mention at SIGMOD 2025 in Berlin.
• EDBT 2025 Best Demo Award: We were also thrilled to pick up a Best Demo Award at EDBT 2025 in Barcelona for “Virtual: Compressing Data Lake Files”.

Collaborations

Research in data systems doesn’t happen in an ivory tower. We love working closely with industry leaders and top academic labs to solve real-world problems.

• SIGMOD 2025 Industrial: We presented “Pruning in Snowflake: Working Smarter, Not Harder” in collaboration with Snowflake, where we explored some cool new pruning techniques.
• Google BigQuery & SemBench: Our work on SemBench, a benchmark for semantic query processing engines, is a huge team effort with Google BigQuery, Cornell, TU Berlin, University of Michigan, MIT CSAIL, and Vrije Universiteit Amsterdam.

A Benchmark-Heavy Year

Benchmarks are the foundation of systems research. While we’ve been involved in benchmarking for a while (you might know our past work on JOB-light, SOSD, and Redset), we’ve ramped up our efforts in 2025:

• Redbench: We presented it at aiDM 2025, and we’re continuing to develop this workload synthesis work in collaboration with TU Darmstadt.
• SemBench: Our new benchmark, designed specifically for evaluating semantic SQL operators.

Community Building and Interdisciplinary Research

Beyond the core research, we’ve been connecting with the broader community:

• Bavarian Database Day: We’re proud to have organized the very first Bavarian Database Day, which brought together researchers and practitioners from Bavaria and beyond.
• Podcast Feature: Mihail Stoian hopped on the Disseminate podcast to talk about our work on robust query execution in the episode on Parachute.
• UTN Internal Milestone: We even published a first cross-department workshop paper, which was a big step for interdisciplinary research here at UTN.

Everything Else!

It wouldn’t be a proper recap without mentioning all the other cool stuff we did:

• BTW 2025: I was co-organizing the ML4Sys and Sys4ML workshop at BTW 2025, and I also gave a talk about “Workload-Driven Indexing in the Cloud”.
• Dagstuhl Seminar: I attended a very cool Dagstuhl seminar on Table Representation Learning, which is actually where we first kicked off our work on SemBench.
• Lab Offsite: We had a nice summer offsite south of Nuremberg. Between working on ML-based data compression, we managed to squeeze in some fun swimming and wakeboarding at the nearby lake.
• VLDB Presentations: We were also busy in London at VLDB. We presented “Instance-Optimized String Fingerprints” at AIDB and Parachute at the main Research Track.
• Long Night of Sciences: We presented our agentic data analytics platform DataLoom and our data lake file format Virtual at the Long Night of Sciences.

All in all, it’s been a crazy year, and I couldn’t be more proud of what we’ve achieved. I’m super excited for what 2026 has in store. Stay tuned!