Linux

We are back in the c-base! Do you already run NixOS and want to talk to other people or want an introductory session with some NixOS users. Just bring your machine! Usually there are no topics, we just sit in front of our computers, talk about stuff, debug some problems or review some PRs on nixpkgs. There are drinks usually available from the c-base bar. If you can't find us. just ask around or ask in irc (#krebs on hackint)

**Chat & Updates:** Telegram: [https://t.me/+VfHrJXPGVWLflGhz](https://t.me/+VfHrJXPGVWLflGhz) Discord: [https://discord.gg/SVb44D9](https://discord.gg/SVb44D9) **\*Deutscher Text unten\*** **Come & go as it fits your schedule** We are meeting in-person and online every week. We offer an office environment for self-study, work and presentation/discussion (lightning talks). **Schedule:** * 10am open doors * 10am online video conference open [https://meet.in-berlin.de/fcc](https://meet.in-berlin.de/fcc) * optional, afternoon, around 1pm: 1h presentation session (if requested / offered) * 6pm or extended open-end **Open co-working / co-learning** work on the FreeCodeCamp curriculum or your own project. Vibe inside but vape outside. **Mutual exchange** and support is welcome. English and German, as needed. We have fast internet, chocolate bars, drinks and a nice atmosphere. :-) **Personal Consultation / Career advice** Bodo runs this meetup for many years. As a Senior Dev/DevOp he is open to giving guidance for your questions about career development. Book a free appointment (up to 60min) with him here: [https://calendar.app.google/yoZfVZR4D9zwt5in7](https://calendar.app.google/yoZfVZR4D9zwt5in7) **Code presentations on the projector** Everybody can participate. Show your code and present your project/problem. Get feedback or help and let others learn from you! You don't have to prepare a fancy presentation, show us what you're working on. Also in English or German, as you feel more comfortable. **Komm und gehe jederzeit. Zeitplan:** * 10:00 Uhr Öffnung vor Ort * 10:00 Uhr online room [https://meet.in-berlin.de/fcc](https://meet.in-berlin.de/fcc) * Nachmittag, ca 13 Uhr: optional 1 Std. Präsentation / Gruppen-Diskussion (nach Angebot/Bedarf) * 18:00 bzw. länger, open-end **Präsentationsrunde** nach Bedarf und Angebot am Nachmittag. Dort man kurz zeigen, woran man gerade arbeitet, eventuell Hilfe bekommen oder interessante Tool vorstellen. Du musst keine große Präsentation vorbereiten. Ebenfalls auf Englisch oder Deutsch, wie Du Dich wohl fühlst. **Gegenseitiger Austausch** und Unterstützung sind willkommen. Es wird Englisch und Deutsch gesprochen. **Persönliche Beratung / Karriereempfehlung** Bodo veranstaltet dieses Meetup seit vielen Jahren. Als Senior Dev/DevOp gibt er Dir auch persönliche Ratschläge für Karriere im 1:1 Gespräch. Reserviere Deinen kostenfreien Online-Videocall (bis zu 60min) hier: [https://calendar.app.google/yoZfVZR4D9zwt5in7](https://calendar.app.google/yoZfVZR4D9zwt5in7) **Tech & Buzz:** Claude Code, nodeJS, npm, nvm, TypeScript, JavaScript, Bootstrap, Tailwind, Bulma, React, Angular, NextJS, Vue, Nuxt, Docker, Container, Virtualization, Django, Codex, LLM, vibe coding, static site generation, OAuth2, Simple SAML, UX, UI, CMS, markdown, Hugo, Netlify, freeCodeCamp, Udemy, Community, Frontend Masters, Backend, SQL, noSQL, MySQL, Python, Linux, Development, git, Version Control, HTML, CSS, Open Source, Responsive Web design, Relational Database, API, Data Science, Machine Learning, Infosec, Cybersecurity, Big Data, AI, Artificial Intelligence, ChatGPT, Stable Diffusion, Data Structures, Data Mining, Ethical Hacking, OSI, Network Security, Julia, Deno, Test driven development, CI, continuous delivery, Deployment Pipe, Bash, devOps, Gitlab, GitHub, Jira, Cloud Computing, Azure, AWS, Apache Server, SSH, Algorithm optimization, SEO, C/C++, Mobile Apps, Apple iPhone, mac OS, Firefox, Chrome, Android, Programming Languages, Cryptography, Patching, Blockchain, DApp, Solidity, Analytics, geek, nerd, Neural Network, MongoDB, ExpressJS, Flutter, Go Lang, AJAX, OOP, UML Diagram, Computer Science, Engineering, Modell, View, Controller, Requirements, risks analysis, PKI, SSL, PGP Certs

**Details** Join us at Bödikerstraße 25 for a yet another delightful DevOps meet-up, this time hosted by Enpal! **Agenda** 18:00 Open Doors 18:25 Enpal Introduction 18:30 Talk 1 - From ClickOps to Scalable Platform *by Pepijn Schoen* 19:00 Food and Beer 19:30 Talk 2 - It works… but no one understands it: The hidden problem in DevOps *by Dayana Mick* 20:00 Networking & Drinks 21:00 Close ➡️ Interested in speaking at this event? Fill out our [Call for Speakers](https://forms.gle/51NCoVixa3Us4Crz7) -> please hand in your talks for this event until 07.04. latest THX ➡️ Interested in hosting an event? Fill out our [Call for Hosts](https://docs.google.com/forms/d/e/1FAIpQLScBpga8Yp-QAQt2lGskZWOThaKyQ_MbMbZaNcklgeXxqr5Vrg/viewform?usp=sf_link) and let's set up a meeting. *Pepijn Schoen* **From ClickOps to Scalable Platform** For engineers and engineering leaders, on culture and decision making. Abstract We ran everything in one Azure subscription. Leads and sales, installations and steering energy systems. Everyone had access. Most resources were created manually. It was unclear who owned what. As the tech org grew from 100 to 300 people, reinventing security and scaling individually stopped working. We'll cover how we approached platform building as gardeners rather than architects: observing what already works, replicating it, and letting rituals emerge before encoding them. We'll look at some mistakes we made along the way and what we learned from each. *Dayana Mick* **It works… but no one understands it: The hidden problem in DevOps** For several years, I kept asking the same question to senior developers, staff engineers, and mentors: “What is a build?” I rarely got a clear answer. Not because people didn’t know, but because the mental model had faded somewhere between Terraform, Ansible, Kubernetes, and “just run this command.” At some point, I realized I could run pipelines, deploy services, and fix things just enough to move forward, but explaining what was actually happening was much harder. And I was not the only one. This talk is about a pattern I’ve seen across every company I’ve worked at. Systems that work, but are not really understood. Teams running platform commands from memory. Platform teams becoming bottlenecks. Incidents where even experienced engineers hesitate. And situations where people rely on tools without a clear sense of what is actually happening underneath. With AI, this is only accelerating. This is a talk for anyone who has ever copy-pasted a command, watched it work, and decided not to ask further questions. It is especially relevant for junior and mid-level engineers navigating DevOps complexity, and for teams who are trying to scale tooling without losing shared understanding.

We are gathering the robotics community again in a new, exciting location! Join us for the **Berlin Robotics Meetup** on **April 21**st at **Erwin Schrödinger-Zentrum of HU Berlin**! We’ll feature an exciting **demo (RoboCup)** from the Adaptive Systems Group, the Computer Science Department of Humboldt University, as well as additional talks and presentations from the community. More details about the talks and additional speakers will be announced soon – stay tuned! 🗓️ When? April 21st (Tuesday) 📍 Where? Erwin Schrödinger-Zentrum, Rudower Chaussee 26, Hall 0'119, 12489 Berlin (S Adlershof) ### Confirmed Speakers: **Stella Alice Schlotter** @ *Adaptive Systems Group* **Paul Beck-Danilov** @ *NormaCore* **Anton Maltsev** @ Rembrain.ai ... More speakers to be confirmed ### Agenda 6:00–6:25 pm – Doors open 6:25–6:30 pm – Welcome note 6:30–8:00 pm - Talks and Q&A 8:00–10:00 pm – Networking & Robot Demo No online stream is planned. ### Interested in Speaking? If you’d like to present your project or product at a future meetup, reach out to us via LinkedIn. Spots are limited – register now on Meetup and secure your place! Whether you’re a seasoned engineer or just starting your journey, come be part of the community that’s building the future of Berlin robotics together!

Join this in-person workshop in Berlin hosted by Mozilla.ai to build an AI agent with Octonous agent platform. ​No technical skills required. Just bring your laptop and your curiosity. **YOU MUST RSVP HERE:** **[https://luma.com/octonous-berlin-2026](https://luma.com/octonous-berlin-2026)** You won't automatically be accepted if you register on Meetup: ​**💻 About the event** ​This is an in-person, interactive product session in Berlin designed for professionals in: * ​Operations * ​Customer Support * ​Sales * ​Marketing * ​HR * ​Product ​Especially those who regularly deal with repetitive workflows and constant tool switching. ​If you've ever thought: * ​"Why am I still doing this manually?" * ​"There must be a smarter way to handle this." * ​"AI is powerful, but how would it actually work for my job?" ​This session is for you. ​You won't just watch a demo. You'll test the product yourself in an avant-premiere. Your feedback will directly shape how Octonous evolves ahead of its beta launch. **​Spots are limited to keep the session interactive.** We give priority to participants who can bring a real workflow they wish to automate. RSVP but **you must apply for a ticket here:** **[https://luma.com/octonous-berlin-2026](https://luma.com/octonous-berlin-2026)** \-\-\-\-\-\-\- For the latest Berlin Tech News, Events & Jobs head to **[techinberlin.com](https://www.techinberlin.com/)**

We're back and excited to feature **Nouamane Tazi**, who is currently a Research Engineer at **HuggingFace** and will discuss "**Scaling LLM Training to Thousands of GPUs**", lasting approximately 45 minutes. After the talk, seize the opportunity to connect with fellow AI enthusiasts to share ideas and questions while enjoying free drinks and pizza. **Door close by 7.15pm, so please come early! Also, "attend"ing (RSVP) here on Meetup is strictly necessary to be guaranteed entry.** Please note that Meetup has recently been quite keen on promoting its Plus program. However, you are not obligated to purchase it, as both our events and the platform remain free. **Who is this event for?** This event is open to everyone interested in state-of-the-art AI research. We especially design it for students, PhD candidates, academic researchers, and industry professionals with a research focus in machine learning. **Abstract**: Training large language models at scale introduces a cascade of systems bottlenecks absent at smaller scales: from communication overhead and memory fragmentation to subtle numerical instabilities that surface only across thousands of devices. This talk covers the practical design choices behind scaling LLM training to thousands of GPUs: what parallelism strategies work (and when they break), how to keep training runs efficient and stable, and the engineering trade-offs that shape modern pretraining infrastructure. The presentation aims to be accessible to a broad ML audience, drawing on real-world experience from large-scale open-source training runs at Hugging Face. **Bio**: Nouamane Tazi is a Machine Learning Research Engineer at Hugging Face, specializing in training and scaling large language models. He is a co-author of SmolLM3 and The Ultra Scale Playbook, and his research spans NLP, deep learning, and scalable AI infrastructure. We are [BLISS](https://bliss.berlin/) e.V., the **AI** **organization** in Berlin that **connects like-minded individuals** who share great interest and passion for the field of machine learning. This summer 2026, we will, again, host an exciting speaker series on site in Berlin, featuring excellent researchers from cohere, ETH Zürich, University of Oxford, HuggingFace, and Stanford University. Website: [https://bliss.berlin](https://bliss.berlin/) Youtube: [https://www.youtube.com/@bliss.ev.berlin](https://www.youtube.com/@bliss.ev.berlin) Disclaimer: By attending this event you agree to be photographed.

 Gemeinsam kochen & genießen Hast du Lust, neue pflanzliche Rezepte kennenzulernen und in entspannter Atmosphäre gemeinsam zu kochen? Am 21. April 2026 um 18:30 Uhr laden wir dich ins Offene Wohnzimmer ein, um zusammen drei leckere Gerichte zuzubereiten: – Kichererbsen-Kokus-Curry – Linsensalat mit Zitronen-Tahini-Dressing – Pfannengemüse mit Kräuterdip Egal ob vegan, neugierig oder einfach hungrig – komm vorbei, lerne neue Menschen kennen und entdecke, wie einfach und lecker pflanzliche Küche sein kann.  Ein großes Dankeschön an das Offene Wohnzimmer für die Möglichkeit, diesen Abend bei euch zu gestalten! 

Hello fellow Symfonians, we are thrilled to announce the next user group meeting of the year, featuring two insightful talks that you **won't want to miss**! This time we are hosted by c-base (Rungestrasse 20 10179 Berlin). Come by and enjoy an evening of learning, networking, and socializing with fellow Symfony and PHP developers. **Agenda:** 18:30: Doors open 19:00: Welcome and Introduction 19:20: **Talk 1: "Maintainable Frontends for Symfony" by Daniela Berger** 19:50: Break & Snack 20:00: **Talk 2: "Why every CMS needs MCP" by Roland Golla** 20:40: Socializing **Talk Details:** **Talk 1: "Maintainable Frontends for Symfony" by Daniela Berger** In many web projects we find more backend devs than frontend devs, sometimes significantly so. This often leads to devs with little frontend experience being asked to do frontend development anyway, and they often end up shaping their frontends in ways that will make their lives unnecessarily hard. One common result are frontends that are functionally unmaintainable because each change - whether it is a new feature or a bugfix - requires implementing an exception to existing code, thus making the code increasingly and unnecessarily complex. This is especially upsetting because with a bit of experience it is possible to implement light-weight frontend components with equally elegant CSS and JavaScript that are easy to extend and to adapt to new requirements without digging oneself into a deeper hole with each PR. And since Symfony is making frontend development steadily more accessible to backend devs with Symfony US / Stimulus, this is a good time to take a closer look at \*how\* to structure these frontends. This talk aims to a) introduce core frontend concepts that backend devs might not be aware of, b) show best frontend practices both on the code level and on the architecture level, c) show a few common mistakes that can be found in inexpertly implemented frontends, and c) introduce tools that will make consistent frontend development easier. **Talk 2: "Why every CMS needs MCP" by Roland Golla** Your content team already works in ChatGPT, Claude, and Gemini. Yet the content still lands in the CMS backend manually, clumsily, slowly. MCP changes that. One sentence in chat becomes a published article. No copy pasting, no formatting, no clicking. And the content ranks. On Google, on ChatGPT, on Perplexity, on whatever comes next. Good content written by AI directly into the CMS goes live faster and gets found. I show you the MCP plugin for Sulu CMS: open source, built on Symfony, ready to deploy. But this is not about implementation. It is about the three questions every content team must ask: Why does every website need MCP? What does good AI content look like? And why is conversational content management better than anything you click together in a backend? MCP makes content creation as easy as chat. If you can write, you can publish. No CMS training, no workflow, no waiting. Live on stage: a complete workflow from idea in chat to published article in Sulu CMS. Everything open source on GitHub. Don't miss these insightful talks, engaging discussions, and networking opportunities. We can't wait to see you at the **April Symfony User Group**!" If you have any questions or accessibility requirements, please reach out to us. Also, if your company wants to be the next host for the User Group, just let us know!

All are welcome. Whether you are new to Berlin, have lived here for a long time or are just passing through. Come on a Thursday, enjoy a drink and have a good time, meet new people and have interesting conversations with us. The place is wonderful, the first cocktail bar in Prenzlauer Berg. The offer only for Meetup members is Aperol Spritz for 6 € the whole evening. Now takes place on Thursday again. :)

Wichtig: Bitte meldet euch über den [offiziellen Link der Symfony Group](https://www.meetup.com/de-de/sfugberlin/events/314105549/?eventOrigin=group_upcoming_events) an Hallo liebe Community, die Symfony User Group trifft sich bald wieder – und ich freue mich besonders darauf, dieses Mal selbst einen Talk beizusteuern. Außerdem steht eine spannende Session mit Daniela Berger auf dem Plan, die euch garantiert neue Impulse für euren Symfony‑ und PHP‑Alltag liefert. Das Treffen findet bei c-base statt (Rungestraße 20, 10179 Berlin). Wer Lust auf einen Abend voller Austausch, frischen Ideen und netten Gesprächen hat, ist herzlich eingeladen vorbeizuschauen. **Programm:** 18:30 Uhr: Einlass 19:00 Uhr: Begrüßung und Einführung 19:20 Uhr: Vortrag 1: „Wartbare Frontends für Symfony“ von Daniela Berger 19:50 Uhr: Pause & Snack 20:00 Uhr: Vortrag 2: „Warum jedes CMS MCP braucht“ von Roland Golla 20:40 Uhr: Geselliges Beisammensein **Vortrag 1:** **„Wartbare Frontends für Symfony“ von Daniela Berger** In vielen Webprojekten gibt es mehr Backend- als Frontend-Entwickler, manchmal sogar deutlich mehr. Dies führt oft dazu, dass Entwickler mit wenig Frontend-Erfahrung trotzdem mit der Frontend-Entwicklung betraut werden, und sie gestalten ihre Frontends häufig so, dass sie sich das Leben unnötig schwer machen. Ein häufiges Ergebnis sind Frontends, die funktional nicht wartbar sind, da jede Änderung – sei es eine neue Funktion oder ein Bugfix – die Implementierung einer Ausnahme vom bestehenden Code erfordert, wodurch der Code zunehmend und unnötig komplex wird. Das ist besonders ärgerlich, da es mit ein wenig Erfahrung möglich ist, schlanke Frontend-Komponenten mit ebenso elegantem CSS und JavaScript zu implementieren, die sich leicht erweitern und an neue Anforderungen anpassen lassen, ohne sich mit jedem PR in ein tieferes Loch zu graben. Und da Symfony mit Symfony US / Stimulus die Frontend-Entwicklung für Backend-Entwickler immer zugänglicher macht, ist jetzt ein guter Zeitpunkt, sich genauer anzuschauen, \*wie\* man diese Frontends strukturieren sollte. Dieser Vortrag zielt darauf ab, a) zentrale Frontend-Konzepte vorzustellen, die Backend-Entwicklern möglicherweise nicht bekannt sind, b) bewährte Frontend-Praktiken sowohl auf Code- als auch auf Architekturebene aufzuzeigen, c) einige häufige Fehler aufzuzeigen, die in unsachgemäß implementierten Frontends zu finden sind, und d) Tools vorzustellen, die eine konsistente Frontend-Entwicklung erleichtern. **Vortrag 2:** **„Warum jedes CMS MCP braucht“ von Roland Golla** Ihr Content-Team arbeitet bereits mit ChatGPT, Claude und Gemini. Dennoch landet der Content immer noch manuell, umständlich und langsam im CMS-Backend. MCP ändert das. Ein Satz im Chat wird zu einem veröffentlichten Artikel. Kein Kopieren und Einfügen, keine Formatierung, kein Klicken. Und der Content rankt. Bei Google, bei ChatGPT, bei Perplexity, bei was auch immer als Nächstes kommt. Guter Content, der von KI direkt ins CMS geschrieben wird, geht schneller online und wird gefunden. Ich zeige Ihnen das MCP-Plugin für Sulu CMS: Open Source, auf Symfony aufgebaut, bereit zur Bereitstellung. Aber hier geht es nicht um die Implementierung. Es geht um die drei Fragen, die sich jedes Content-Team stellen muss: Warum braucht jede Website MCP? Wie sieht guter KI-Content aus? Und warum ist dialogbasiertes Content-Management besser als alles, was Sie im Backend zusammenklicken? Mit MCP ist das Erstellen von Inhalten so einfach wie Chatten. Wer schreiben kann, kann auch veröffentlichen. Keine CMS-Schulung, kein Workflow, keine Wartezeiten. Live auf der Bühne: ein kompletter Workflow von der Idee im Chat bis zum veröffentlichten Artikel im Sulu CMS. Alles als Open Source auf GitHub.

**🏆 Win a free ticket to [DASH26](https://dash.datadoghq.com/)!** We’re hosting an on-site raffle where the grand prize is a ticket to Datadog’s annual conference in New York City. \-\-\-\- All talks will be **presented in English**, to ensure that as many people as possible can participate and engage at this event. **If you want to attend, please RSVP to secure your spot - this will make organizing easier. Thank you so much ♥️** **Location:** [The-B Berlin, Revaler Str. 32, 10245 Berlin](https://www.theb-berlin.com/) \-\-\-\-\-\- **🏠 18:00 - Arrival: Networking, Drinks & Snacks (30 min.)** Grab yourself snacks & drinks and say hello to everybody else! **📅 18:30 - Introduction & What's new at Datadog? (15 min.)** **🎙️**Speaker: Marcel Drechsler, User Group Leader & Product Owner Internal Developer Platform @ **[andsafe](https://andsafe.de)** Introduction into the evening and highlights of Datadog's recent new features and products. **📅 18:45 - Powering Platform Engineering through Datadog (30 min.)** **🎙️**Speaker: Marcel Drechsler, User Group Leader & Product Owner Internal Developer Platform @ **[andsafe](https://andsafe.de)** In the rapidly changing landscape of Platform Engineering, Datadog has evolved from a monitoring tool into a comprehensive foundation for Internal Developer Platforms (IDPs). This session explores the journey of scaling observability and security into a unified platform strategy that reduces developer friction. We will dive into how Datadog’s expanding ecosystem provides the essential building blocks for modern self-service infrastructure. Attendees will learn how to leverage these integrated features to build a more resilient and transparent developer experience. Discover how to transform your Datadog instance into a strategic asset for your platform’s success. **📅 19:15 - Logs as a First-Class Citizen - How Lightspeed Commerce evolved logs to unlock the full power of Datadog** **(30 min.)** **🎙️**Speaker: Rein Martha, Staff Software Engineer, Lightspeed Commerce When Lightspeed started with Datadog, we didn't begin with traces or metrics — we began with evolving our logs. Raw, unstructured, and full of noise. The first step was making them worth keeping: trimming duplicates, removing what no one ever read, and transforming what remained into structured, queryable signals. That foundation changed everything. Once logs became first-class — with clean attributes, consistent structure, and a clear purpose for every line — the rest of the observability stack followed naturally. Monitors built on log queries. Dashboards that actually meant something on incidents. Metrics generated directly from log attributes, giving us long-term retention without the cost of keeping everything raw. **📅 19:45 - Best Practices for Alerting with Datadog (30 min.)** **🎙️**Speaker: Santiago Gomez Saez, Datadog ambassador & Principal Cloud Architect @ **[dxone](https://www.dx.one.gmbh/)** Operational excellence is the main objective of SRE teams. Focusing on alerting, this talk shares common pitfalls and best practices on how and when to alert when incidents occur. In addition, we show how to self-heal in some cases requiring no manual intervention. **🥗 20:15 - Drinks, Food & Networking** Enjoy refreshments while networking with community peers! **👋 21:00 - Goodbye, see you next time!**

Join us in Berlin for an evening dedicated to the present and future of mobile development with Mendix. In the first session, a developer from BlueBerk will share real-world insights from a recently completed large-scale Mendix Native project. The talk will cover the current state of native development, common limitations teams face today, and practical workarounds that proved successful in production. In the second session, a Product Manager from Mendix will provide a forward-looking perspective on mobile development with the Mendix platform. He will discuss the available mobile options today and what’s coming next, helping you understand how to choose the right approach for your project. The presentation will compare native mobile and Progressive Web App (PWA) strategies and introduce the PWA Wrapper as a flexible and practical alternative for many scenarios. Whether you are currently building mobile apps or planning your next Mendix project, this meetup will give you valuable hands-on experience, architectural guidance, and a clear view of the platform’s mobile

Hola chicos, Moin zusammen, lasst uns Deutsch und Spanisch sprechen und ein paar Bier zusammen trinken, die Idee ist einfach und gut :) Bitte bringt zumindest basis Deutsch oder Spanisch Kenntnisse mit! Vamos a hablar español y alemán y tomar unas cervezas, la idea es simple y buena jaja. ¡Por favor, ven por lo menos con un nivel básico de español o alemán! Danke, gracias!

Join our in-person meetup on April 24th to hear talks from experts on cutting-edge topics across AI, ML, and computer vision. **[Register to reserve your seat.](https://voxel51.com/events/berlin-ai-ml-and-computer-vision-meetup-april-24-2026)** Space is limited! **Date, Time and Location** Apr 24, 2026 5:30 PM - 8:30 PM [MotionLab](https://motionlab.berlin/) Bouchéstraße 12/Halle 20 12435 Berlin **Kaputt: A Large-Scale Dataset for Visual Defect Detection** We present a novel large-scale dataset for defect detection in a logistics setting. Recent work on industrial anomaly detection has primarily focused on manufacturing scenarios with highly controlled poses and a limited number of object categories. Existing benchmarks like MVTec-AD (Bergmann et al., 2021) and VisA (Zou et al., 2022) have reached saturation, with state-of-the-art methods achieving up to 99.9% AUROC scores. In contrast to manufacturing, anomaly detection in retail logistics faces new challenges, particularly in the diversity and variability of object pose and appearance. Leading anomaly detection methods fall short when applied to this new setting. To bridge this gap, we introduce a new benchmark that overcomes the current limitations of existing datasets. With over 230,000 images (and more than 29,000 defective instances), it is 40 times larger than MVTec and contains more than 48,000 distinct objects. To validate the difficulty of the problem, we conduct an extensive evaluation of multiple state-of-the-art anomaly detection methods, demonstrating that they do not surpass 56.96% AUROC on our dataset. Further qualitative analysis confirms that existing methods struggle to leverage normal samples under heavy pose and appearance variation. With our large-scale dataset, we set a new benchmark and encourage future research towards solving this challenging problem in retail logistics anomaly detection. The dataset is available for download under [https://www.kaputt-dataset.com](https://www.kaputt-dataset.com). *About the Speaker* [Sebastian Höfer](https://www.linkedin.com/in/sebastian-h%C3%B6fer-891178121/) is an Applied Science Manager at Amazon Fulfillment Technologies & Robotics, leading machine learning and computer vision research for large-scale robotics and warehouse automation. He received his PhD from the Robotics & Biology Lab at TU Berlin, focusing on Sim2Real transfer and robotic perception. His recent work, “Kaputt: A Large-Scale Dataset for Visual Defect Detection” (ICCV 2025) [37], established a major benchmark for industrial anomaly detection, reflecting his expertise at the intersection of academic research and real-world deployment. **Data Foundations for Vision-Language-Action Models** Model architectures get the papers, but data decides whether robots actually work. This talk introduces VLAs from a data-centric perspective: what makes robot datasets fundamentally different from image classification or video understanding, how the field is organizing its data (Open X-Embodiment, LeRobot, RLDS), and what evaluation benchmarks actually measure. We'll examine the unique challenges such as temporal structure, proprioceptive signals, and heterogeneity in embodiment, and discuss why addressing them matters more than the next architectural innovation. *About the Speaker* [Harpreet Sahota](https://www.linkedin.com/in/harpreetsahota204/) is a hacker-in-residence and machine learning engineer with a passion for deep learning and generative AI. He’s got a deep interest in VLMs, Visual Agents, Document AI, and Physical AI. **Most AI Agents Are Broken. Let’s Fix That** AI agents are having a moment, but most of them are little more than fragile prototypes that break under pressure. Together, we’ll explore why so many agentic systems fail in practice, and how to fix that with real engineering principles. In this talk, you’ll learn how to build agents that are modular, observable, and ready for production. If you’re tired of shiny agent demos that don't deliver, this talk is your blueprint for building agents that actually work. *About the Speaker* [Bilge Yücel](https://www.linkedin.com/in/bilge-yucel/) is a Senior Developer Relations Engineer at deepset, helping developers build agentic AI apps with Haystack. Passionate about AI, she makes complex concepts approachable through hands-on tutorials, both online and at real-life events. **Operationalizing Computer Vision for Overhead Lines: Beyond the Demo** At first glance, visual inspection of high-voltage power lines seems straightforward: collect imagery, run one or two AI models, and report the findings. In practice, moving beyond a proof of concept reveals a range of issues that can make or break a campaign. Common concerns include data quality and coverage, scarcity of the most relevant cases and abundance everywhere else, variations in pylon geometry and asset types across regions, calibration and GIS alignment challenges, and a long tail of edge cases that emerge in real-world operations. This talk introduces Siemens Energy’s end-to-end overhead line inspection solution and shares key learnings from inspecting more than 10,000 km of power lines for real customers across several continents. We will show how raw 2D/3D data is transformed into structured information, delivering insights into asset inventory as well as defects, and supporting maintenance and planning decisions for critical infrastructure. The focus is on the combination of algorithmic building blocks and scalable processing, designed for robustness and consistency at scale, where even low error rates can become operationally significant. *About the Speaker* [Stefan Wakolbinger](https://www.linkedin.com/in/stefan-wakolbinger-aa0ba874/) is the Development Team Lead for AI & Analytics at SIEAERO, Siemens Energy's digital powerline inspection service. He leads the development of cutting-edge AI and analytics solutions that transform aerial powerline inspection through multi-sensor technology. His team creates digital twins of powerline infrastructure, automates fault detection, and monitors vegetation management—making powerline inspection safer, more precise, and more efficient. Stefan has been driving innovation in this role since September 2022. **Search your video library like a database** Drop in YouTube URLs or upload files and query content four ways: exact keyword matching, semantic search across transcripts, visual scene search via SigLIP2, and LLM-generated answers that synthesise across segments. [Paras Mehta](https://www.linkedin.com/in/pmehtaeu/) is a Berlin-based AI engineer and CTO/co-founder of Sylby, a language learning app he built from scratch, reaching 10,000 users and raising €350K. Previously: data scientist at Motionlogic, senior software engineer at Volkswagen, a PhD from Freie Universität Berlin, and a visiting stint at Cambridge. He now works as an AI engineer at HPI's AI Service Centre.

Bring your work or your hobby, hang out, and code with us. Follow @buckeyecocoa for more information.

**💥 Join us THIS Saturday and EVERY Saturday!** **💥 At Columbus’ Hottest New Salsa Venue!** **.** Lincoln Street Salsa THIS Saturday and EVERY Saturday! 8 pm – 1 am **.** **💥 Gorgeously Renovated Studio!** **💥 Free Parking!** **💥 BYOB!** **💥 Easy Access from Highway!** **.** Lincoln Street Salsa 1717 Brice Rd Reynoldsburg, Ohio 43215 **.** **ENTRANCE IS ON THE NORTH SIDE OF THE BUILDING** **Behind Anthony’s Pharmacy!** **.** **PLEASE SHARE and help get the word out!** **.** **🟣 Salsa, Bachata, and Merengue Dancing!** **.** **🟣** 8:00 pm – 8:30 pm **𝗥𝗲𝗴𝗶𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻** **.** **🟣** 8:30 pm – 9:30 pm **𝗕𝗲𝗴𝗶𝗻𝗻𝗲𝗿 𝗗𝗮𝗻𝗰𝗲 Class!** **.** **🟣** 9:30 pm – 1 am **𝗢𝗽𝗲𝗻 𝗗𝗮𝗻𝗰𝗲 𝗣𝗮𝗿𝘁𝘆!** **.** **✅ Questions?** **Visit [LincolnStreetSalsa.com](https://columbussalsadancing.com/)** Or Text Todd: 614-774-8146 **.** **✅ No Experience Necessary!** **✅ No Partner Required!** **✅ Plenty of People to Dance With!** **.** **✅ $15 Cash / $16 Card** **✅ College Students $5 Off with Valid ID!** **✅ BYOB** **✅ Age 18+** **.** **✅ Lincoln Street Salsa is the perfect place for…** **.** **👉 Learning to Dance** **👉** Social Dancing **👉 Birthday Celebration** **👉** Bachelorette Parties **👉 Girl’s Night Out** **👉 Date Night** **👉** Getting Fit / Burning Calories **👉 Meeting New People** **👉** Having Fun, Fun, Fun! **.** ✅ **Questions?** Visit [LincolnStreetSalsa.com](https://columbussalsadancing.com/ "https://columbussalsadancing.com/") Or Text Todd: 614-774-8146 .

Want to be a speaker? submit your talk to our Call for Presenters!!! https://sessionize.com/cbus-hug-2026/

We practice, dance, and enjoy a BYOB bottle of your favorite wine at the Dublin Ballroom off Sawmill, North, and “milonga” is a social event for tango dancing where dancers socialize. A “*practica*” is an informal social event for tango dancing where dancers focus on practicing the dance; while a “*practilonga*” combines the relaxed atmosphere of a “*practica*” with the format of a “milonga.” We are doing it all! ***Cost is $10 payable to the Teacher.*** *Cost is for the cost of the venue and covers some snacks.* We have had fun doing this event in the past. Please arrive promptly as lessons begin at 2pm. Don't worry about bringing a partner as there will be others at the event from outside the group that: a) might have a clue and b) won't step on your toes. The event is BYOB (similar to our Salsa events), so bring your favorite and a wine glass. We will likely find a place for dinner and drinks after the dance.

**Improving Office in Franklinton** Physical location: Improving Office 330 Rush Alley Suite #150 Columbus, OH 43215 Schedule: 6:00 p.m.: Socialize, eat, and drink. Improving will be providing pizza and beverages. 6:30 to 8:00 pm. Main meeting and presentation(s). Topic: This month Chris Pazsint will be talking about Agentic Coding. How does one use CLI Based Agents, and Agentic IDEs such as Cursor, Kiro, Antigravity? How to include agentic coding plugins for IDEs you already love such as Visual Studio Code. We meet on the last Monday of each Month. Presentations are given by members and friends of this group. If you would like to do a presentation (small or large) on a python topic, please contact Central OH Python at centralohpython@gmail.com

**Important time note:** Please plan on arriving between 5:30 and 6:00 as the elevators lock after 6 and you'll need to message us and we'll need to come get you. The building address is 4450 Bridge Park The entrance is 6620 Mooney St, Suite 400 You will need to scan your ID at the door to get a visitor badge. **Abstract** GitHub Copilot is powerful, but what if you could scale from a solo AI assistant to an entire team of specialized agents working in parallel? This session introduces Squad: an open-source framework for multi-agent orchestration that lets you define teams of AI agents with specific roles, responsibilities, and expertise. We'll progress from Copilot basics to the Copilot CLI, explore how Agents add autonomy, and see how Instructions and Skills let you customize agent behavior. Then, the climax: a live demo where a Squad team of 3 agents (Lead, Developer, Tester) stands up and builds a working application in real-time, showcasing true multi-agent collaboration. Whether you're new to AI or exploring how to scale your use of Copilot, this session will show you what's possible when agents work as a team. **YouTube Link** TBD