Congratulations to Sebastian Baunsgaard successfully defending his PhD thesis, "Workload-aware Compressed Linear Algebra for Data-centric Machine Learning Pipelines" Sebastian is part of the DAMS group (lead: @matthiasboehm7.bsky.social) BIFOLD / @tuberlin.bsky.social #phd #defended

Meet the BIFOLD researchers at #SIGMOD2025 in Berlin! From June 22–27, we’re excited to welcome the global data community and contribute with papers, demos & workshops—plus help shape the event as part of the organizing team. See you there! 🚀 #datamanagement News www.bifold.berlin/news-events/...

Successfully defended my PhD yesterday! 🎓 🎉 Special thanks to my mentor Sebastian Möller and professors Sina Zarrieß @clausebielefeld.bsky.social, Christin Seifert, and @matthiasboehm7.bsky.social for being part of my committee. Will continue working on XAI & NLP as a post-doc at TU Berlin & BIFOLD

Here is an additional impression from BTW'25 in Bamberg. I'm terrible at such interviews, but not hiding anything. 😉 Thanks to the organizers for a great conference. www.youtube.com/watch?v=qSx4...

A huge thank you to our speakers and everyone who contributed to the great discussions at today’s workshop! A big shout-out to Manisha Luthra and @matthiasboehm7.bsky.social for co-organizing—always a pleasure working together! 😊

A postcard from Lausanne - I just returned from the Applied Machine Learning Days (AMLD) at EPFL. Thanks to Thaleia Dimitra Doudali and Pamela Delgado for organizing the "AI in Data and Computer Systems" session - it was a pleasure.

Together with Riccardo Tommasini, we are excited to announce Yuanyuan Tian & Juan Reutter as the keynotes for the Transforming Graph Data Workshop@EDBT/ICDT 2025! 🌟 Join us March 25 in Barcelona 🇪🇸 👉 transforming-graphs-workshop.org 👉 edbticdt2025.upc.edu?contents=reg... cc: @rictomm.bsky.social

Please help spread the word by reposting! We've just created the official DEEM Workshop account: @deem-workshop.bsky.social

📢📢VLDB26 (@vldb.bsky.social) will have a ShadowPC! Apply at forms.office.com/e/XYihKj4UKq by March 1! Great opportunity for folk who are new to the data management community & want to gain paper reviewing experience. More info at application form. Chairs: @zistvan.bsky.social, Tianzheng Wang, & I.

🎥 Review key insights from The Age of AI: Panel Discussion with Sam Altman #OpenAI, Nicole Büttner #MerantixMomentum, Fatma Deniz @tuberlin.bsky.social, and Volker Markl @bifold.berlin discussed AI's opportunities, challenges, and societal impact. www.bifold.berlin/news-events/...

Thanks to Sam, Nicole, Volker, and Fatma for the great panel discussion at TU Berlin & @bifold.berlin today.

BIFOLD RSG Lead Matthias Boehm will be a speaker at the "AI in Data and Computer Systems" track of the AMLD - Applied Machine Learning Days at EPFL. @appliedmldays.bsky.social @tuberlin.bsky.social @matthiasboehm7.bsky.social #AMLDEPFL2025 #DataSystems #ComputerSystems #Innovation t1p.de/iqhx8

I'm very proud of my team - we just completed three great BTW'25 papers (camera-ready versions and reproducibility packages for two of them) as well as the final EU project review meeting of DAPHNE today: btw2025.gi.de github.com/damslab/repr... github.com/daphne-eu/da...

Our graduate school has an open call for 10 PhD positions in data management and machine learning (until Feb 03). This is a great opportunity to join one of our groups, sometimes even with co-supervision among multiple groups.

Many thanks to our guest Gerardo Vitagliano for his great talk on “Making Data Accessible using LLMs.” 👏 Thanks for the great talk! @csail.mit.edu @tuberlin.bsky.social @ziamurai.bsky.social @mersault.bsky.social @matthiasboehm7.bsky.social

The @sigmod2025.bsky.social Programming Contest goes into another round. We (Bo Tang, Tilmann Rabl, and myself) just published the timeline and task overview: sigmod-contest-2025.github.io/index.html Thanks to Carlo Curino and @microsoft.com for the continued support.

Proud advisor moment: Today, my first PhD student @arnabphani.bsky.social successfully defended his PhD thesis in front of a great committee of Ana Klimovic, Tilmann Rabl, me, and @mersault.bsky.social (w/ summa cum laude - very good with distinction). Arnab is on the job market, so don't miss out.

There is a special place in hell for schools that require you to rate applicants across 10 axes for recommendation letters and don't just let you upload the standard PDF.

I'm happy to promote the K.I.E.Z. (Künstliche Intelligenz Entrepreneurship Zentrum, linkedin.com/company/kiez...), where I'm also part of the expert network. Startups can get support for Apr-Sep 2025. Bridge-to-Market: lnkd.in/dSzDTY-n Accelerator: lnkd.in/dE_WWTct

Happy NeurIPS Poster Session is Full day to all who celebrate

We are excited to be starting a new lab that rethinks the systems, ML, and HCI stack to support agent-based automation! Columbia's PhD deadline is 12/15 -- please share with potential applicants! columbia-dap-lab.github.io

We now extended the deadline of the ML4Sys/Sys4ML@BTW workshop to Dec 15 luthramanisha.github.io/ML4Sys-and-S...

Happy to share this paper, which was in the making for a while (@arnabphani.bsky.social‬ showed great perseverance and continuously improved the prototype, experiments, and paper presentation).

We've shipped a very neat performance improvement at Materialize, which I'll try and unpack but you can get a tl;dr by just looking at the memory decrease in the changelog announcement: materialize.com/changelog/20... Big memory reductions for "temporal filters". What are those? Let me explain.

We've been developing a serverless runtime for Transactional Cloud Apps. Here is why we adopted (streaming) stateful dataflow graphs. "Transactional Cloud Applications Go with the (Data)Flow" K. Psarakis, G. Christodoulou, M. Fragkoulis, A. Katsifodimos. CIDR'25: kpsarakis.com/assets/pdf/s...

We're excited to announce that the VLDB Summer School will take place from Monday, 28th of July, to Thursday, 31st of July, 2025 at the University of Wroclaw, in Poland! 🇵🇱 More details to follow soon ☀️ www.linkedin.com/company/vldb...

The deadline for industrial papers is on the same day. Submit your most innovative industrial data systems work there: 2025.sigmod.org/calls_indust...

Vol:17 No:12 → MLOS in Action: Bridging the Gap Between Experimentation and Auto-Tuning in the Cloud 👥 Authors: Brian Kroth, Sergiy Matusevych, Rana Alotaibi, Yiwen Zhu, Anja Gruenheid, Yuanyuan Tian 📄 PDF: https://www.vldb.org/pvldb/vol17/p4269-kroth.pdf

🚀 Up to 93x input compression for LLMs! By compressing the data in the KV cache, we squeeze more info in the context. Presented at @emnlpmeeting.bsky.social, now on MIT Press: FINCH: Prompt-guided Key-Value Cache Compression for LLMs (TACL 2024) direct.mit.edu/tacl/article... More details 👇

At cwi.nl/en/events/dijkstra-awards/cwi-lectures-dijkstra-fellowship we awarded the Dijkstra Fellowship to Marcin Zukowski for his contributions to DB architecture in a full room & great speakers: Allison Lee, Viktor Leis, @andypavlo.bsky.social @hannes.muehleisen.org @madelonhulsebos.bsky.social

There are less than two weeks until the deadline of our ML4Sys & Sys4ML workshop @ BTW2025 in Bamberg. luthramanisha.github.io/ML4Sys-and-S...

📯📯📯 Round 3 accepted papers for SIGMOD 2025 are now posted on the website. Congratulations to all the authors and see you at the conference in Berlin! 2025.sigmod.org/sigmod_paper...

Now that "scale is all we need" has predictably faltered... 1. Small AI models often perform better in context. 2. Obsession w bigness has bad consequences, from climate, to power concentration, to research capture. Me + @gaelvaroquaux.bsky.social @sashamtl.bsky.social arxiv.org/abs/2409.14160

My PVLDB paper announcement bot now supports Bluesky: @pvldb.bsky.social I am letting it post papers from last issue (August 2024). It will then hit the new volume when it pops later this year. You can also get RSS/Atom feeds: db.cs.cmu.edu/files/rss/pv... db.cs.cmu.edu/files/rss/pv...

btw, last week I also completed my first "mental health first aid" 2-day course and hope to be able to help students when needed in the future.

something to get your fired up ("inspired") for the next week: www.youtube.com/watch?v=WwPS...

Welcome new followers and check out the Mystery AI Hype Theater 3000 starter pack!

Why I love DuckDB in one anecdote: I needed to copy all tables from one database to another, and I was sure it was going to take a while to figure it out. I googled "copy from one duckdb database to another," and this was the first result:

Pls repost: We, the DEEM Lab at TU Berlin, are hiring a postdoctoral researcher in data engineering for machine learning. Details available at: deem.berlin#jobs-57624 This fully-funded position is part of the Berlin Institute for the Foundations of Learning and Data (BIFOLD). #databs #datasky

The world doesn’t live on a pixel grid and neither should vision models! Excited to share Moving off-the-Grid (MooG): a video model w/o grid-based representations. MooG learns detached “off-the-grid tokens” that bind to (and track) scene elements as camera & content move. 🧵

Last Wednesday, we had a great Queens Lecture by Aldo Faisal at TU Berlin (I'm at the bottom right), followed by a lovely dinner at the residence of the British Ambassador.

𝗔 𝗽𝗲𝗲𝗸 𝘂𝗻𝗱𝗲𝗿 𝘁𝗵𝗲 𝗵𝗼𝗼𝗱 𝗼𝗳 𝗗𝘂𝗰𝗸𝗗𝗕 If you want to know what makes DuckDB so fast, check out the new blog post by Tim Ebergen on how the "optimizers play a silent, but vital role when using a database": duckdb.org/2024/11/14/o...