Johannes Kolbe's talk at #bbuzz investigates Client-Side AI using WebGPU, ONNX Runtime, and Transformers.js. It also explores the reality of hardware access, model size, and the 2026 trade-offs of browser-based execution. Slides: 2026.berlinbuzzwords... Watch the recording: youtu.be/v1sCCqwFYNk...
Berlin Buzzwords
Berlin Buzzwords
@berlinbuzzwords.de
Berlin Buzzwords is Germany's most exciting conference on storing, processing, streaming and searching large amounts of digital data, with a focus on open-source software projects. Next edition: 13-15 June 2027 https://berlinbuzzwords.de/
62 followers22 following211 posts
Emilie Ma's talk shows how agents can enable widespread formal methods cheaply and reliably, with valuable results. Slides: 2026.berlinbuzzwords... Recap: youtu.be/M2wb2ug2gYg...
In this panel, moderated by Monica Sarbu, speakers share what agentic workloads actually look like in production: what broke, what had to be rebuilt, and where the architecture is heading. Check out the recording: youtu.be/LjpQz2K3ahM...
Hidden bias in datasets silently breaks machine learning systems in production. Stas Don presented practical techniques for detecting hidden bias in datasets before models reach production in his talk. Slide: 2026.berlinbuzzwords... Watch the full recording: youtu.be/O2Zl_shq6JM...
In this talk by Zoi Kaoudi and Haralampos Gavriilidis, discover how Apache Wayang - a recently graduated Apache Top Level Project provides a unified framework for cross-platform execution. Download the slide: 2026.berlinbuzzwords... Full recap: youtu.be/yrjg3a7CiBo...
When does Apache Flink solve real problems versus add complexity? Naci Simsek was at #bbuzz to show use cases where Flink becomes essential such as fraud detection, CDC, real-time analytics versus when batch or Kafka Streams suffice. 2026.berlinbuzzwords... Recording: youtu.be/RqgytnmlW7k...
At #bbuzz 20026, Jarek Potiuk joined us to break down how the Apache Software Foundation is thinking about Open Source Maintainership in the age of Agentic AI, and why AI's biggest contribution isn't writing code, it's giving maintainer time back. Watch the recording: youtu.be/gsm4UA-azAA...
Zoi Kaoudi and Haralampos Gavriilidis joined us at #bbuzz26 to introduce Apache Wayang, a recently graduated Apache Top Level Project that lets you write a pipeline once and run it efficiently across Spark, Flink, JDBC databases, and ML systems. Full recording: youtu.be/yrjg3a7CiBo...
Ralph Matthias Debusmann was at #bbuzz26 to break down why existing stream processing libraries are too complex, and to unveil Kafi Streams, an open source Python library that treats streaming like batch. Full recording: youtu.be/_oKQaDyP1CQ...
Streaming Agents can invoke LLMs directly from Flink for each event, but for high-volume Kafka topics that gets expensive fast. Steffen Hoellinger showed how combining this with Complex Event Processing in Flink cuts cost and enforces deterministic behavior. Recording: youtu.be/AyNluIol8Ho...
Did people feel scared when AI first stepped into the spotlight? Charlie Hull shares his own first encounter with it, recorded live at #bbuzz 2026. He argues AI is a new challenge for search, not a replacement for it or the people behind it. Watch the episode: youtu.be/m8D0iBo0UK4
At #bbuzz 2026, David Louis Hollembaek and Vincent Pistor looked through 20 years of search evolution to show how knowledge graphs and vector search fix the missing link in today's Agentic AI. youtu.be/6r9-Wf-NkI8...
Many novel ML techniques started as hacks that happened to work, but the demands of building real systems can be at odds with this creative culture. William Benton was at #bbuzz 2026 to argue that ML practitioners and infrastructure teams aren't as different as they appear. youtu.be/Q-DiIh7x1jU...
Postgres, Apache Kafka, and Apache Iceberg are three popular data services, and none of them is quite like the other, especially when it comes to writes. At #bbuzz 2026, Celeste Horgan from Snowflake compared writes across all three in his talk. Recording: youtu.be/8yb9XfDRQSA...
Ankit Jain joined us at #bbuzz 2026 to explore how columnar data layouts, SIMD-accelerated execution, and bulk-oriented processing are reshaping search internals, in the talk "From Inverted Index to Columnar Vectorized Execution Search." Full recording: youtu.be/5HxDnUQprXE...
How much CO2 is your cloud usage responsible for? Dmitry Kan asked Julien Nioche, founder of DigitalPebble and creator of the open-source tool SPRUCE, recorded live at #bbuzz 2026. They cover what SPRUCE is and how it gets deployed. Watch: youtu.be/-IY7s00gxTk...
Apache Solr 9.11 brings in-place index upgrades with zero downtime. No more recreating large indexes from source on major version upgrades. At #bbuzz 2026, Rahul Goswami covered it in "Zero downtime index upgrade in Apache Solr." Watch the recording: youtu.be/6fJVrx11Nbw...
Filip Makraduli's talk covered why hybrid retrieval requires multiple model types, the adapter architecture that makes multi-model serving possible, and building the pipeline end-to-end, tradeoffs, and lessons. Watch the recap: youtube.com/watch?v=... #bbuzz
Radu Gheorghe and Rafał Kuć were at Berlin Buzzwords 2026 to talk about preparations for the golden set and untangling the dependencies with the title "Circular Dependency Fixes when Bootstrapping a Golden Set". Check out the recording: youtube.com/watch?v=... #bbuzz
This talk, "DuckDB beyond the notebook" by Matthias Niehoff at Berlin Buzzwords 2026 shows how DuckDB changes the way we build data-driven applications and also explored scenarios including browser-based analytics powered by WebAssembly. Full recording: youtube.com/watch?v=... #bbuzz
Is search obsolete in the age of AI? Dmitry Kan sat down with Trey Grainger and Doug Turnbull, co-authors of "AI-Powered Search," talking about how AI is actually changing search, recorded live at #bbuzz 2026. Watch "Role of Search in Modern AI": youtu.be/acOGVynTVpM
Hartmut Armbruster during his talk at Berlin Buzzwords 2026 "What If We've Been Scaling Stream Processing Wrong All Along", shared an early-stage exploration of a framework that preserves the Kafka Streams DSL and so much more. Sounds exciting? Full recording: youtube.com/watch?v=... #bbuzz
Stanislav Don was at Berlin Buzzwords 2026 with "Detecting Hidden Bias in Datasets Before Models Fail" to A show us how to detect data imbalance, leakage, and coverage gaps early before misleading offline metrics turn into costly real-world failures. Watch it here: youtube.com/watch?v=... #bbuzz
At Berlin Buzzwords 2026, Gülçin Yıldırım Jelinek joined us with "What you should know about constraints in PostgreSQL 18" and gave a detailed walkthrough of the pg_constraint catalog, what’s new in PostgreSQL 18 and what’s coming in PostgreSQL 19. Full session: youtube.com/watch?v=... #bbuzz
Dmitriy Kostunin and Julian von Hoerschelmann-Schliwinski joined us at Berlin Buzzwords 2026 with "AI in the physical world: from observation to discovery" and shared real-world applications of AI across modern physics and more. Catch the live recording here: youtube.com/watch?v=... #bbuzz
At Berlin Buzzwords 2026, Apurva Misra joined us with this session "How to Tell If Your Agent Used the Right Stuff" and explored ways to tell whether your agent truly saw and used the right context. The full recording: youtube.com/watch?v=... #bbuzz
Berlin Buzzwords 2026 wrapped up over a month ago, and preparations for #bbuzz 2027 are already underway! We're heading back to Kulturbrauerei, Berlin for our 18th edition — mark your calendars for 13–15 June 2027. Full details: 2026.berlinbuzzwords...
Berlin Buzzwords kicked off with a keynote from Ruth Suehle: "Building Resilience: The Next Decade of Open Source." A session that explored ways open source has evolved Watch the recording: youtu.be/-7qO37l2YfI... #bbuzz
That's a wrap for Berlin Buzzwords 2026! Thank you to everyone who joined us for the seventeenth edition of #bbuzz. Three days of talks, demos, and conversations across search, data, and AI at Kulturbrauerei. Check out our blog post recap of the conference: 2026.berlinbuzzwords...
A big thank you to our Gold Partners OpenSearch Project, Qdrant, and Vespa.ai, our Recording Partner Bloomberg, our Silver Partners Hornet.dev, Linkup, NetApp Instaclustr, Thomson Reuters Lab, and Search Guard, our Bronze Partner OpenSource Connections.
Also, thank you to our Get-Together Partner Aiven, our Community Breakfast Partner Xata.io, and our Contributor Partner AI-Powered Search. See the full list of partners here: 2026.berlinbuzzwords.de/our-partners/
That's Day 1 of Berlin Buzzwords 2026 done! Thank you to everyone who joined us today across all four stages and online. See you tomorrow for Day 2! Check what's on: 2026.berlinbuzzwords... #bbuzz
Looking to power real-time AI applications on your data at any scale? Connect with the team at the Vespa.ai booth! 🤖 ⚡️ #bbuzz
The first day of #bbuzz is in full swing! Time to take a well-deserved coffee break to recharge your batteries ☕🧁. A huge thank you to our partner Vespa.ai for sponsoring the break! Go grab one and let us know: which cupcake flavor is your favorite? 👇
Nearly a year after #bbuzz25, it’s almost that time again. #bbuzz26 is just around the corner, here’s a look back at last year’s edition before we do it all over again. Join us for Berlin Buzzwords on June 7-9 June 2026 Kulturbrauerei or online 2026.berlinbuzzwords...
Final reminder for #bbuzz26 The full schedule is live to help you plan your days, map out your favorite sessions, and make the most of your time in Berlin Buzzwords. 2026.berlinbuzzwords...
1 day to go until #bbuzz26. 🚀 It’s almost time. See you tomorrow at Berlin Buzzwords, and definitely at the Barcamp 🥳 2026.berlinbuzzwords...
Before we head into #bbuzz26 weekend, here’s a look back at #bbuzz24. Get your tickets now: 2026.berlinbuzzwords...
Join the Community Breakfast for Women* on Monday at Berlin Buzzwords, made possible by Xata. Connect, exchange ideas and start the day together. Spots are limited, RSVP here: 2026.berlinbuzzwords...