AI & ML Research Events in Europe

We are working through Build a Reasoning Model (From Scratch) by Sebastian Raschka.
During each meetup, we will discuss the key concepts from the chapter or pages under review. In this session we'll discuss the Chapter 6. Training Reasoning Models with Reinforcement Learning.

Raschka provides a hands-on guide to understanding and building modern reasoning models. Starting with a conventional pretrained LLM, the book explores evaluation, inference-time reasoning techniques, reinforcement learning with verifiable rewards, and knowledge distillation.

Please note that the session is not recorded, and participants are responsible for obtaining their own copy of the book.

Buy the book (affiliate links):
Amazon UK
Amazon France

Book overview:
In Build a Reasoning Model (From Scratch), bestselling author Sebastian Raschka explains how modern reasoning-oriented language models work by implementing their core techniques step by step.
You will begin with a conventional pretrained LLM and learn how to generate and evaluate its responses. You will then explore inference-time techniques such as chain-of-thought prompting, sampling, self-consistency, response scoring, Best-of-N, and self-refinement. Later chapters introduce training-based approaches, including reinforcement learning with verifiable rewards, GRPO, format rewards, and distillation from stronger reasoning models into smaller ones.
Build a Reasoning Model (From Scratch) teaches you how to:

• Implement core LLM reasoning techniques from scratch
• Generate, score, and evaluate model responses
• Build verifier-based evaluation systems
• Improve reasoning using self-consistency, Best-of-N, and self-refinement
• Apply reinforcement learning with verifiable rewards
• Understand and implement GRPO-based training
• Distill the reasoning capabilities of stronger models into smaller ones
• Evaluate the accuracy, cost, and latency trade-offs of different reasoning techniques

Ever wondered how AI actually works?

ATU Sligo Room B2314

Open to all ATU Sligo students and staff

Spend two hours building one yourself. You'll write real Python code, train a model to recognise handwritten digits, and test where it succeeds and fails — no coding experience required, no software to install. A hands-on introduction to the technology shaping science, medicine, and research..

Embedded AI v praxi: Bezpečnost, doprava a vesmírné technologie

Zveme vás na další setkání Embedded AI komunity v Brně, které tentokrát propojí umělou inteligenci, kybernetickou bezpečnost, embedded vývoj a dopravní a vesmírné technologie.

Setkáme se 10. listopadu 2026 od 17:00 v prostorách Centra VUT Tržnice v Brně.

Akce se koná pod záštitou FEKT VUT a VUT Tržnice a je součástí festivalu Dny AI 2026.

Čekají vás přednášky, praktické ukázky, krátké lightning talks i prostor pro otevřenou diskusi a networking.

Přijďte se inspirovat, sdílet vlastní zkušenosti a poznat další lidi, které zajímá propojení AI, hardwaru a embedded systémů.

Na co se můžete těšit?

Přednášky:

✅ Hackeři s AI: Jak umělá inteligence mění útoky na IoT a embedded systémy

Jak se mění kybernetické útoky, když útočníci začnou využívat umělou inteligenci? Diana Průšová z Yunex Traffic ukáže, jak může AI pomoci při analýze zranitelností, identifikaci attack vektorů a přípravě útoků na IoT a embedded systémy.
Na praktických příkladech představí, jak AI mění možnosti útočníků, jaké nové příležitosti jim otevírá a co to znamená pro bezpečnost IoT a embedded zařízení. Zaměří se také na konkrétní důsledky pro vývojáře a na to, jak na tyto hrozby reagovat.

✅ VZLUSAT-2: CubeSat s Linuxem a prvními krůčky k AI

Tomáš Novotný z VZLU AEROSPACE, a.s. představí nanosatelit VZLUSAT-2, který zahájil svou téměř čtyřletou vesmírnou cestu v roce 2022 a jehož hlavním cílem bylo pořizování snímků zemského povrchu a sběr vědeckých dat z měření gama a rentgenového záření. Na palubě družice byly dva linuxové počítače postavené na komerčně dostupném systému na čipu. Použití takových počítačů na nízké oběžné dráze Země vyžaduje zvýšenou odolnost vůči radiaci a teplotním výkyvům v prostředí s omezenými možnostmi chlazení. Použití Linuxu umožnilo snadnou automatizaci palubních procesů a rozšíření funkcionalit systému, ke kterým patřila i klasifikace obrazových dat pomocí neuronové sítě na palubě družice.

✅ AI při vývoji bezpečnostního čipu

Jan Medek a Saadya Chevan z Tropic Square ukáží, jak jejich tým využívá AI při vývoji bezpečnostního čipu. Zaměří se na jednotlivé fáze vývoje, agentní pracovní postupy, human-in-the-loop kontrolu, CI gates i situace, ve kterých AI stále selhává.
Řeč bude také o bezpečnosti modelových zařízení Edge AI, například kamer nebo průmyslových bran, a o tématech, jako jsou secure boot, podepsané aktualizace, bezpečné ukládání klíčů nebo vzdálené ověřování zařízení.

Praktické ukázky a dema:

Vedle přednášek si budete moci prohlédnout několik konkrétních řešení:

• awareAI 2.0 od Yunex Traffic – hardware pro detekci dopravních informací z kamer s využitím neuronových sítí a platformy NVIDIA Jetson. Součástí ukázky bude také cloudový AI chatbot určený ke vzdálené správě a monitoringu zařízení.
• KOSTKA od YSpace – model první české studentské družice v měřítku 1:1, živá data ze satelitu a možnost setkat se přímo s členy vývojového týmu.
• Springboard od Edhouse – platforma pro tvorbu digitálních produktů a propojování systémů s lidmi, využitelná například pro strojové vidění, prediktivní údržbu nebo řízení průmyslových procesů.
• VZLUSAT-2 od VZLU AEROSPACE, a.s. – model družice v měřítku 1:1 doplněný o záznamy telemetrie ze sítě SatNOGS.
• Demo Chip-to-Cloud od Tropic Square – ukázka bezpečnostního čipu v praxi a diskuse nad bezpečností zařízení, identitou zařízení, bezpečnými aktualizacemi a požadavky Cyber Resilience Act**.**

Máte vlastní projekt nebo zkušenost?

Součástí programu budou také krátké lightning talks. Pokud máte zajímavý projekt, praktickou zkušenost, technologii nebo problém z oblasti Embedded AI, o který byste se chtěli podělit, přihlaste svůj krátký příspěvek při registraci.
Nemusí jít o velkou prezentaci. Stačí několik minut, konkrétní zkušenost a chuť otevřít zajímavé téma.

Proč přijít?

Chceme vytvořit otevřený prostor pro sdílení zkušeností z oblasti Embedded a Edge AI. Baví nás práce s platformami jako Nvidia Jetson, ARM, Intel i ESP, hrajeme si s embedded OS (Yocto, Ubuntu), ladíme kód v C, C++ i MicroPythonu a sledujeme nové technologie včetně powerline komunikace.

Prosíme všechny účastníky, aby při registraci uvedli:

•Jméno a příjmení
•Společnost

Kapacita akce je omezená. Organizátor si vyhrazuje právo rozhodnout o účasti jednotlivých účastníků v případě překročení kapacity nebo z organizačních důvodů.

Těšíme se na vás!

Gabka Golcová z Yunex Traffic za organizační tým Brno Embedded AI Community

How can we uncover the most meaningful factors hidden in complex data and understand what each of them actually represents?

We are happy to welcome Prof. Dr. Ullrich Köthe from Heidelberg University to heidelberg.ai to talk about EOFlows, a non-linear extension of Principal Component Analysis (PCA).
PCA identifies the main factors of variation in a dataset, but it is limited to linear relationships. Normalizing flows can capture more complex patterns, but they lack interpretability.
EOFlows aim to combine the flexibility of normalizing flows with the interpretability of PCA. Using orthogonality regularization, they learn latent dimensions that capture distinct factors of variation, which can be ranked by their explained entropy.
In this talk, Prof. Köthe will explain how EOFlows work and show what they discover in the CelebA portrait dataset. He will also discuss how stable the results are across training runs and how the method compares with existing approaches.

About Ullrich Köthe
Ullrich Köthe is an Associate Professor at Heidelberg University, where he has led a group in the Visual Learning Lab since 2018. His research spans image analysis, machine learning, and explainable AI, with a focus on developing methods that make the inner workings of neural networks easier to understand.

For more info go to: https://www.heidelberg.ai/2026/11/10/ullrich-koethe.html

After the event, we will also upload a video recording to our YouTube channel: https://www.youtube.com/channel/UCfHWBneOsb7SfOxJepnMQKA

Visual retrieval is moving beyond text. Join us for two technical talks: teaching models to search documents by treating pages as images, including the problem of imperfect labels in training, and using intent detection to build image discovery in a fashion AI assistant. Pizza and drinks on us — stick around after for networking.
✍️ RSVP soon — capacity is limited!

📅 Date and Time
Wednesday, November 11, 2026 from 18:00-20:45

📍 Location: Hedwig-Wachenheim-Straße 7, 10243 Berlin
BHW Auditorium (1st floor), Zalando SE
📝 Agenda
18:00 – Doors open, arrivals and check-in
18:15 – Welcoming and open session
18:30 – Teaching Models to Search Documents: What If the “Wrong” Answer Is Right? by Saba Sturua, Senior AI Research Engineer @ Elastic
How do we teach models to search PDFs with tables, charts, and complex layouts? I’ll introduce visual document retrieval and explore how larger models can help train smaller search models, including why useful answers can be mistaken for wrong ones during training. This work is part of the research behind our next generation of Jina embedding models.
19:00 – Q&A and quick break
19:15 – From Conversation to Visual Search: Building Image Discovery for Zalando AI Assistant by Inanda Menezes, Principal Software Engineer @ Zalando
How can a fashion AI assistant understand a customer's intent from visual input? In this talk, we’ll explore how we built image discovery for Zalando’s Assistant, considering the intent detection to support different use cases, from finding similar products to getting fashion advice. We’ll dive into how we connect these intents to the right processing and retrieval strategies, and share the key challenges.
19:45 – Q&A, pizza, drinks & networking
***
Invite your friends and join us for an evening of learning and networking with fellow engineers and speakers!

Want to speak at our next meetup? Submit your talk on this link and we will get in touch with you as soon as possible.

Join our virtual meetup to hear talks from experts on cutting-edge topics across AI, ML, and computer vision.

Time, Date and Location

Nov 11, 2026
9:00 AM - 11:00 AM PST
Online. Register for the Zoom!

Agentic RAG: Beyond Retrieve-and-Generate

Retrieval-Augmented Generation (RAG) has become the default pattern for grounding large language models in external knowledge, but most implementations still follow a rigid retrieve-once, generate-once pipeline — one that struggles with multi-hop questions, ambiguous queries, and knowing when its own retrieved context is insufficient. This talk introduces Agentic RAG, where retrieval is treated as an action within an agent's reasoning loop rather than a fixed upstream step.

We'll examine query decomposition and routing for breaking complex questions into targeted sub-queries, self-reflective and corrective retrieval loops that let an agent judge and re-query its own results, and tool-orchestration patterns (via MCP) that let retrieval sit alongside other agent actions like database lookups and API calls. Using a live architecture — evolving a standard RAG chatbot into an agentic, MCP-connected system — we'll walk through what changes in design, and where these systems introduce new failure modes: grounding drift, latency and cost from repeated retrieval loops, and cases where a simpler RAG pipeline still wins.

Attendees will leave with a practical framework for deciding when the added complexity of agentic RAG is worth it, and a set of design patterns for building it correctly.

About the Speaker

Balaji Venkatasubramaniyar is a Technical Lead at Wisdom Infotech, leading a 15+ person engineering team delivering enterprise solutions. With 13+ years of experience in enterprise software and insurance technology, he specializes in agentic AI systems, RAG architectures, and vector databases.

GeoAI for the Physical World: Earth Observation, Foundation Models, and Urban Digital Twins

Earth observation provides a unique form of computer vision for understanding the physical world at city to continental scales. In this talk, I will show how satellite imagery, geospatial data, machine learning, and foundation-model representations can be combined to characterize urban environments and environmental conditions.

I will present UrbanScope Open, an open GeoAI digital-twin prototype integrating Earth observation with 3D buildings, vegetation, land-surface temperature, air quality, noise, population, and other urban data. I will also share lessons from my research using geospatial foundation-model embeddings for environmental prediction across Europe.

The talk will discuss how these approaches can contribute to increasingly multimodal AI systems capable of reasoning about real-world environments.

About the Speaker

Cesar Alvarez is a researcher at the University of Augsburg working at the intersection of GeoAI, Earth observation, remote sensing, and environmental intelligence. His research applies machine learning, computer vision, and geospatial foundation models to problems including urban environments, climate risk, air quality, and agriculture.

Can agents get curious?

Most AI agents are good at answering a question once we tell them exactly what to look for. The harder problem is building agents that can explore a complex dataset autonomously: generating hypotheses, deciding which analyses are worth running, allocating additional compute when evidence is ambiguous, and knowing when they have enough evidence to stop.

In this talk, I’ll show an architecture for autonomous research agents that combines structured knowledge, iterative tool use, and explicit evidence tracking to turn open-ended questions into a sequence of testable investigations. I’ll discuss practical lessons from building and evaluating these systems, including why more test-time compute does not automatically produce better research and how provenance and evaluation can make long-running agents more reliable.

I’ll close with a live example of an agent exploring a dataset, revising its hypotheses, and choosing what to investigate next.

About the Speaker

Srivatsa P is a member of technical staff at Sigma Computing, where he works on AI agents that reason over complex enterprise data. Previously, he worked on machine learning at Apple and conducted research at the MIT Media Lab; outside of traditional ML, he has also worked on mapping coral reefs through underwater imaging, which sparked an enduring interest in how intelligent systems make sense of messy real-world data.

Does Cross-Validation Work in Deep Learning?

Cross-validation is a standard in hyperparameter optimization in classical ML, but is generally considered computationally prohibitive and unnecessary in deep learning. We investigated whether CV produces 1. better DL models 2. better estimates of future performance of said models. We focused on medical image classification as an important deep learning application domain.

Our results show that cross-validation consistently and materially reduces the gap between development and independent testing error for relatively small training datasets (up to about 1000 images). The test performance improvement was generally modest. We discuss the implications for deep-learning practitioners, including whether the increased computational cost of cross-validation is justified.

About the Speaker

Ljubomir Buturovic is an AI scientist and biotech executive with extensive experience in clinical diagnostics, including work at Inflammatix and Pathwork Diagnostics. His current research focuses on deep-learning image classification, infectious disease adjudication using LLMs, and development of AI models for helping critically ill patients.

Sponsored by Daemon

PLEASE SIGN UP ON LUMA NOT ON MEETUP!

Agenda
06:00pm – Doors open
06:30pm – Welcome
06:35pm – Shradha Janib, “Traffic sign recognition using deep learning”
07:20pm – Break
07:35pm – David Hughes, Stelia, “Lessons learned from running ML workloads in production”
08:20pm – Lightning talks

• Freeman Madudili, “Deterministic boundaries around AI actions”
• Ozan Dagdeviren, “Why measuring AI skill is harder than it looks”
• 08:50pm – Close, social at The Angel

Please note you will be unable to enter the venue before 6.00pm.
RSVPs will close 24 hours before the event, you may be unable to register after this time but you can still watch online.
If you can’t join us in person you can watch remotely via our YouTube channel.
Our hosts may require that we provide a list of all attendees, please ensure that you register with a name that matches your government issued ID or bank card: if you do not we cannot guarantee you entry to the building.
Please RSVP for the event well in advance if you plan to attend in person and unRSVP if you can no longer attend as limited spaces are available.
Thanks to Civo for hosting and Daemon for sponsoring.

Herzlich willkommen zur betreuten Intelligenz am Vormittag!

Wir treffen uns immer am 2. Donnerstag im Monat frisch ausgeruht von 10 bis 12 in Fürth bei der ABF im GS28:

ABF Gruppe
Gebhardtstraße 28
2. OG Ost (Kaminzimmer)
90762 Fürth
https://gs28.de/

Bitte beachtet die Anfahrtsbeschreibung. Die Parkplätze der Apotheke vor dem GS28 sind tabu! Die Zahl der Plätze ist limiitiert, ihr müsst euch explizit anmelden.

Es warten Kaffee, Internet und Getränke auf euch. Ihr bringt eure Notebooks, Modelle, Datensätze oder einfach nur ganz viele Fragen mit - gemeinsam experimentieren wir mit allem, was wir dabei haben.

Oder wie Markus und Adam gerne so fein sagen: "Doing, not talking!"

Wer möchte, darf zu Beginn gerne kurz ein Thema vorstellen, einen Einblick teilen oder einfach nur seine brennenden Fragen in den Raum stellen.

Der Termin ist so gewählt, dass im Anschluss noch reichlich Luft ist. Wenn es also gerade läuft, dann lassen wir laufen. Der Letzte räumt die Krümel auf :)

Wir freuen uns auf euch!
Max, Thor, Tobias, Stefan² & Steffen

DISCLAIMER: We are experiencing an overwhelming number of subscription requests. We won't be able to accommodate all of them, as we have a strict capacity of 200 seats. We will make a decision on approval around Nov 2 and open a waitlist. Please de-register if you are not coming. Thank you for your understanding.

We will prioritize founders -> investors -> corporates -> policy -> academics, striking a balance for representation of all groups.

Day 1 of the ChemAI conference.

Artificial Intelligence is driving a fundamental shift in chemistry and materials science — from intuition-led research to predictive design and increasingly autonomous laboratories.

We invite you to an exclusive gathering of top voices and actual builders discussing scaling businesses, industrial innovation, and strategic investments. 30% Vision, 70% Execution.

Hear from:

• Prof. Max Welling - CTO at CuspAI

• Prof. Steven Latré - VP of AI at IMEC

• Dr. Greg de Temmerman - Director Deeptech at TechLeap

• Dr. Dmitri Jarnikov - Sr Director Data Science at Prosus

• Dr. Aaike van Vugt - CEO at VSParticle

• Dr. Maksym Plakhotnyuk - CEO at ATLANT 3D

• Dr. Maryam Ghadrdan - CEO at Planck Technologies

• Dr. Stefano Martiniani - NYU, co-founder of Matterstack.ai

• Dr. Hanne Falsig - Scientist at TOPSOE

• Investor Panel by Hello Tomorrow with

• Yoram Wijngaarde - Founder & CEO Dealroom

• Johan Stins - Sr Investment Manager Invest-NL

• Johannes Triebs - PartnerU2V VC

In partnership with Techleap, Hello Tomorrow, TTT-AI, Startup Amsterdam, Amsterdam Chemistry Network, and AmsterdamAI.

If you’re joining us for ChemAI Deep Tech Builders, why stop there? Join us for the Conference on Nov 13 www.acnetwork.nl/chemai

Note: When approaching the event capacity (200), we might prioritize those who already have a ticket for Day 2.

ChemAI was founded as an independent, international platform to connect technology leaders, industry innovators, and top academics shaping this transformation. It is where AI meets real-world chemical and materials discovery, offering you real OMG-moments. Check the winter Amsterdam edition, and the summer New York edition, at www.chemai.world

The model was never the hard part. The systems, the data, and the people around it were.

Join us on Thursday, 12 November, at the fromScratch Studio HQ, where we’ll hear from a self-described “recovering data scientist*” about what it really takes to get machine learning into production. Through a series of real war stories from deploying ML systems for enterprise clients, we’ll look at where projects actually break down, and what you can do about it. Each story comes with a concrete insight and a practical takeaway you can apply to your own work.

Stavros Dimou from RealMintAI will show us how the engineering best practices built into platforms like Databricks, Fabric, and SageMaker can solve problems that data scientists usually end up fighting alone, and, more importantly, how you can bring those practices into your own infrastructure, whatever platform you use.

And because this is a Practical AI meetup, we won’t stop at theory. In the second half, we’ll go hands-on and live on Databricks, pulling real data from a public API and seeing these ideas in action. You’ll leave with a working repo and setup instructions to run everything yourself on Databricks Free Edition. So, don’t forget to bring your laptop!

About the Speaker
Stavros Dimou is a Machine Learning Engineer at RealMintAI and a Time Series Forecasting specialist, where he architects AI and data engineering solutions for enterprise supply chains. Having led data science production projects both on-prem and on Databricks, he now considers himself a “recovering data scientist,” and this talk draws directly on the war stories that earned him the title.

What to expect
This will be an approximately two-hour session. We’ll spend the first hour on lessons learned from real-world production ML projects, exploring what goes wrong and why solid engineering makes data science projects easier. The rest of the session will be hands-on, with a live walkthrough on Databricks that you can run yourself.

Stick around after the event for beers and networking!

Registration
We will meet from 19:00 to 21:00 at fromScratch HQ (Themistokli Sofouli 70Α). Don’t forget to RSVP to secure your spot!

Huge thanks to our sponsors fromScratch Studio and Noosware for making this event possible.

Please note that all presentations and discussions will be conducted in Greek.

We look forward to seeing you there!

* The ironic moniker “recovering data scientist” was coined by Joe Reis, an industry-recognized data engineer and best-selling author.

Comment transformer des textes en connaissances exploitables ? C'est tout l'enjeu du Natural Language Processing (NLP), ou traitement automatique des langues, la branche de l'intelligence artificielle qui "apprend" aux machines à lire, comprendre et produire du texte. Thibault ROY, Tech Lead AI & NLP chez OppScience, travaille dans ce domaine depuis 20 ans. Il présentera les grandes applications du domaine, puis la rupture provoquée par l'IA générative. Mais cette rupture déplace aussi les difficultés plus qu'elle ne les supprime (coût, confidentialité, hallucinations, etc.). Exemples à l'appui, nous verrons alors comment combiner IA générative, modèles plus légers et règles dans des exemples concrets.

Das Ziel ist, Möglichkeiten darzustellen, die im Sprachraum Java ohne Rückgriff auf Large Language Models eine Klassifikation von Eingangsdaten durchzuführen.
Wobei diese Eingangsdaten nicht Texte sind, sondern reale Ereignisse innerhalb unseres Lebens, Messwerte, Bilder in definiertem Kontext oder Ereigniskombinationen, Kombinationen aus Messwerten oder ähnlichem.
Über die Gewinnung der Daten wird gesprochen.

Referent - Joachim Kröger
Begonnen habe ich in der frühen 80er Jahren auf Mikroprozessoren zu programmieren. So ergab sich ein fast typischer Werdegang: Nach Assembler code für Z80 und Motorola 68000 folgte sofort 'C' als Grundlage für Mustererkennung, als Thema. Das Thema blieb, mit Verfügbarkeit folgte der Paradigmenwechsel auf Objektorientierte Entwicklung mit C++ und mit Verfügbarkeit des Web dann Java. Also von Beginn an. Die Themen wechselten, es kamen Oracle RDBMS hinzu, Tomcat (JEE) CORBA und RMI. Später begann ich eine Lehrtätigkeit für Java über etwa 10 Jahre. Inzwischen seit 20 Jahren selbständig, befinde ich mich gerade mitten in der Produktentwicklung für ein IoT System. 90% Java.

Dive into the world of AI and security hacks with fellow founders and tech pros for a fun, hands-on evening!

Join us in London for a professional networking evening bringing together startup founders, entrepreneurs, cybersecurity specialists, machine learning experts and technology professionals.

Connect with industry peers, exchange ideas, discuss emerging technologies and explore collaboration across cybersecurity, machine learning, AI and digital innovation.

Why Attend?

Build valuable professional relationships. Meet founders, technology specialists and decision-makers from across the industry.

Explore emerging technologies. Discuss developments in cybersecurity, machine learning, artificial intelligence and digital transformation.

Exchange knowledge and expertise. Share experiences, technical insights and ideas with professionals working across the technology sector.

Discover new collaborations. Meet potential partners, clients, investors and specialists with complementary expertise.

Raise your professional profile. Introduce yourself, showcase your expertise and increase your visibility within the technology community.

Stay ahead of industry developments. Discuss new challenges, innovations and opportunities shaping the future of technology.

Who Should Attend?

Startup Founders & Entrepreneurs
Technology startup founders, co-founders, entrepreneurs and business owners developing innovative products and services.

Cybersecurity Professionals
Cybersecurity consultants, information security specialists, ethical hackers, penetration testers and security analysts.

Machine Learning & AI Specialists
Machine learning engineers, AI developers, data scientists, AI researchers and automation specialists.

Technology Executives & Decision-Makers
CTOs, CIOs, IT directors, technology managers and senior digital professionals.

Software Developers & Engineers
Software engineers, application developers, cloud engineers, systems architects and technical specialists.

Technology Consultants & Advisers
Independent consultants, digital transformation advisers, technology strategists and IT specialists.

Investors & Venture Capital Professionals
Technology investors, venture capital firms, angel investors and professionals interested in emerging technology companies.

Cloud & Data Professionals
Cloud architects, data engineers, infrastructure specialists and professionals working with modern technology platforms.

Innovation & Digital Transformation Professionals
Innovation managers, product managers, digital strategists and professionals driving technology development.

Connect. Exchange Ideas. Explore Innovation.

Whether you are building a technology company, developing machine learning solutions, working in cybersecurity or shaping digital innovation, this event offers an opportunity to meet professionals from across the industry.

Book your place and join us in London for an evening of technology, ideas and professional connections.

Workshop für technisch interessierte.

Einlass ab 12:00 Uhr
Beginn 13:00 Uhr

Themen:
- Open Source für Prognosen (mit sktime)
- Offene Foundation-Modelle
- Lokales Hosting von Vorhersagemodellen

Bleibt auch gerne zum Abendvortrag im Anschluss.

Have you ever wondered what goes into making an LLM? Wonder no more, as we will delve into the birds and the bees on how a baby LLM is born and grows into something that can eat the world!

Agenda:
17:00 Arrive, hang out, get some snacks, beer, soda,
17:30 Food!
18:15 Intro
18:30 The Hitchhiker’s Guide to European Portuguese LLMs
Duarte Carmo - Independent AI/ML research engineer and contractor

Large language models are eating the world. Frontier labs keep pushing the boundary, open-weights models are quickly closing the gap, Europe is — as always — stuck somewhere in the middle, and even Portugal has now released AMÁLIA, its own effort in the space.

But what does it actually take to build an LLM trained on a niche dataset? Why would you? And where do you start when there is so little data? In this talk, I’ll walk through my work on European Portuguese LLMs: from building benchmarks, to finding and filtering data, to training models, and everything in between. Even though this talk focuses on Portuguese, it should be easily applicable to any language/domain!

19:30 More socializing, discussions, hanging out
22:00 Doors close

Sponsors:
Thanks to NumFocus for running Pydata!

Special thanks to PROSA for hosting us!

CAIML #45 is going to happen on November 17, 2026, at Lime Connect.

We will have two talks with additional time for networking:

Talk 1: Ana Moya (Hon. Professor of Data Science and BI at ISM) & Pranit Tandom (MSc Business Intelligence & Data Science): Beyond Large Language Models: A Non-Generative Computational Approach to Text Dynamics and Engagement in German Digital News

Understanding what drives audience engagement remains one of the central challenges for digital news organizations. While current discussions are often dominated by Large Language Models and generative AI, this presentation demonstrates how explainable, non-generative computational methods can provide valuable insights into reader behavior and content performance. Rather than focusing on text generation, the presentation highlights how interpretable analytical methods can uncover the linguistic, structural, and emotional characteristics associated with higher engagement. The findings illustrate how news organizations can leverage transparent data science approaches to better understand content performance, support editorial decision-making, and complement emerging AI-driven strategies.

Talk 2: Stefan Christoph (AWS Principal Solutions Architect): Getting Paid by Agents: When AI Becomes the Customer

AI agents are starting to do more than browse and reason, they can also pay. In this talk, Stefan Christoph explores how agentic commerce is becoming real: using HTTP 402 and x402 to let AI agents autonomously discover, evaluate, and pay for content and APIs. We’ll look at what this emerging infrastructure means for developers, publishers, and a web where software itself increasingly becomes the customer.

We will share an agenda soon.

See you in November,
Aaqib, Marc & Fabian

We are excited to feature Jonas Geiping, who is currently a Research Group Leader at Max Planck Institute for Intelligent Systems and will discuss "Safety and Security of LLM Reasoning", 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.

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.

We are BLISS 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 winter 2026/2027, we will, again, host an exciting speaker series on site in Berlin, featuring excellent researchers from Google Deepmind, Anthropic, MPIIS, Imperial College, Microsoft and UCL.

Website: https://bliss.berlin
Youtube: https://www.youtube.com/@bliss.ev.berlin
​Disclaimer: By attending this event you agree to be photographed.

EPOCH Paper Club Vol.III - AI and the future of knowledge

At EPOCH Paper Club, we share a place to nerd out, together as peers, on deeptech and AI. A few bring one paper each on a bigger topic set by Epoch. This time, it's AI and the future of knowledge. They explain it in plain words, then opinions or questions on what they'd challenge. The remainder is open critique.

Who it's for

Curious enthusiasts, and willing to** put the effort into understanding AI.** Students, researchers, professors and people building. You don't need to be an expert. But we ask you to be genuinely inquisitive.

Want to speak?
Fill in this form *by 12th of November *(more detail on presenting and structure):
Open Google Form

The evening

• 18:00 Doors open.
• 18:15 Welcome and ground rules.
• 18:25 Sharing begins.
• 19:30 Wrap-up.
• 19:35 Stay to chat with speakers and peers

Before you come
On Saturday 14th of November we'll send you the papers that will be discussed, together with a short summary of them.

House rules
Ask questions after speakers introduce the big idea, any time. If you have a strong opinion, tell us, and be respectful. Disagree with the idea, not the person. The hosts may ask anyone who isn't respectful to leave.

Good to know
​50 seats. Registration approval is determined on how genuine your answers are. *Don't worry if you are a beginner, this will NOT affect your registration. *We will deprioritize no-shows.

Location:
Echo, TU Delft. We'll send the room number to approved guests.

Coffee, tea and water are provided. Don't forget your cup!

EPOCH films the talks, not the discussion. Please don't record. Photos are welcome at the end of discussions.

About EPOCH
EPOCH is TU Delft's AI Dream Team. We build responsible AI, compete in worldwide competitions, and run AI Cup, the student AI competition. The Paper Club is a series of evenings for everyone who wants to understand AI properly, together.

The first Physical AI Belgrade: a meetup for people who teach robots to work in the real world. Robot learning, teleoperation, human motion data, humanoids.
Agenda

• 17:45 Doors open. A robot arm is already running on the table.
• 18:00 Short intro, then Tihomir Opačić (Orange Hill): What I saw at NVIDIA GTC Berlin. Physical AI, robot learning data and humanoids in 2026: what the companies building robots showed, what people say about training data, and where the industry is heading. With short video reports from the floor.
• 18:40 Second speaker: to be announced soon.
• 19:10 Live SO-101 teleop: a leader–follower robot arm, from human motion to an episode in a robot learning dataset. Try it yourself.
• 19:25 Pizza, drinks and networking (thanks to CDT Hub).
• 20:00 Wrap-up and the date of the next meetup.

Who it's for: engineers, ML and computer vision people, researchers, students and student robotics teams, founders, and anyone curious how robots learn from human motion.
Talks and slides are in English. Free entry, 60 seats. If you can't make it, please cancel your RSVP so someone from the waitlist gets your spot.
The event is photographed and filmed. Teleop episodes are recorded only with your consent and without faces in frame.
Venue: CDT Hub, Kneza Miloša 12, Belgrade, small hall on the first floor. More at https://physicalai.rs

L'AI non si esaurisce nei Large Language Model e non tutti i problemi si risolvono generando testo.

Il 18 novembre, con il nostro meetup "Non solo LLM", allarghiamo lo sguardo ad altri approcci all'AI, con due talk dedicati a modelli che affrontano problemi diversi dalla generazione di linguaggio naturale.

🧠 Decisioni e AI: uno sguardo a Jev
Con Paolo Antinori (IBM) esploreremo i decision model e come utilizzare l'AI per prendere decisioni, non soltanto generare testo.

📊 Foundation Model for Tabular Data
Con Emanuele Fabbiani (TeamSystem) parleremo di foundation model applicati ai dati tabellari, andando oltre le architetture che hanno dominato la recente evoluzione dell'AI.

Due prospettive diverse, un'occasione per approfondire cosa c'è oltre gli LLM e confrontarci con chi lavora concretamente su questi temi.

📅 Programma completo: https://www.improove.tech/events/5882/M16--Non-solo-LLM

🎟️ La partecipazione è gratuita, ma è necessario iscriversi qui: https://www.eventbrite.it/e/biglietti-meetup-16-non-solo-llm-2002990513545

This workshop is a hands-on tour of what actually breaks when you load real LeRobot datasets, and how Voxel51's native LeRobot support in FiftyOne catches it.

Date, Time and Location

Nov 19, 2026
9:00 AM - 10:00 AM PST
Online. Register for the Zoom!

Robot learning data doesn't look like the data that trains most vision-language models. It's temporal, low-dimensional in the parts that matter most, embodiment-specific, and expensive to collect, and increasingly, it's shipped in Hugging Face's LeRobot format. But packaging robot data consistently isn't the same as being able to see it: episodes hide broken metadata, silently dropped sensor modalities, and mismatched schemas between robots that look identical on paper.

Starting from the core differences that make robot data hard (temporal structure, proprioception, action grounding, embodiment heterogeneity, data scarcity), we'll move into live debugging: a metadata bug that broke multiple dataset imports, sensor streams that vanish without an error, and two robots running an identical benchmark whose data is structurally incompatible. By the end, you'll know what to check before you trust a LeRobot dataset, and how to actually look inside one.

Join us for the inaugural meetup of The TPU Community in Krakow!

This is an independent, developer-led space built for machine learning engineers, cloud architects, and systems designers who want to push the boundaries of hardware-accelerated computing.
Our first event features a complete, full-stack storytelling arc-moving from raw hardware and framework integration, up to serving massive models at scale, and finally to building robust, production-grade AI agent systems.
Whether you are already running workloads on Tensor Processing Units (TPUs) or are curious about migrating from GPUs to optimize your compute costs, this is your space to connect, share benchmarks, and learn from real production use cases.

• Date: Thursday, November 19, 2026
• Time: 18:00 – 21:00 CET
• Location: EPAM Hub, Fabryczna 1a, Krakow
• Entry: Free (RSVP required for building security check-in)
• Perks: Cold drinks, pizza, and networking!

• 18:00 – 18:20 | Doors Open
• Arrive early, grab a drink, and meet fellow AI infrastructure enthusiasts.
• 18:20 – 18:30 | Opening: The TPU Community Vision
• Speaker: Magda Kustosz (TPU Community Manager)
• A quick welcome introducing the community’s mission and our exciting roadmap.
• 18:30 – 19:00 | Talk 1: TorchTPU: Native PyTorch Execution on Google TPU
• Speaker: Viacheslav Koval
• The Hardware & Compiler Layer: Unifying PyTorch across TPU and GPU: what does it open up for ML development, and is TPU development possible without Jax?
• 19:00 – 19:30 | Talk 2: Anatomy of an LLM Request: Scaling Self-Hosted Inference with LLM-D
• Speaker: Cagatay Gurturk
• The Infrastructure & Serving Layer: An architectural exploration of what happens to a request when it hits an LLM. Learn how to orchestrate, partition, and scale self-hosted models efficiently on production-grade infrastructure using LLM-D.
• 19:30 – 20:00 | Quick Pizza Recharge & Networking Break (30 mins)
• 20:00 – 20:30 | Talk 3: From Vibing Chaos to Reliable Agents: Harness Engineering in the AI Ecosystem
• Speaker: Darren "Dazbo" Lester (Google Developer Expert)
• The Application & Agentic Layer: Rapid prototyping is fun, but unconstrained "vibe coding" quickly collapses under hallucination and architectural debt. Learn how to shift developer interventions left using Context Engineering, Autonomous Loop Engineering, and Spec-Driven Development (SDD) to ship deterministic, production-ready AI agents.
• 20:30 – 21:00 | Networking
• Chat with the speakers, ask your toughest technical questions, and connect with the community.

Because this event is hosted inside the EPAM Hub, we must share the attendee list with building security in advance.

• Please ensure your Meetup profile displays your real first and last name.
• Registration closes 2 days prior to the event (Tuesday, November 17th, EOD) to finalize the security list.

Can't wait until the event? Join our online spaces to start swapping benchmarks, sharing articles, and troubleshooting compilation errors:
👉 WeAreCommunity Profile: https://wearecommunity.io/communities/tpucommunity
👉 Discord Server: [Link coming soon!]

Let's build the future of accelerated AI together. 🛠️

-------PLEASE REGISTER ONLY THROUGH THIS LINK--------

Join us for another Tech Talk at Glovo HQ 🚀

We’re opening the doors of Yellow Park on Thursday, November 19th, for a Data Science Tech Meetup focused on Real-World Machine Learning at Scale. Together with our peers from Skyscanner, we will explore how we solve some of the most complex challenges in ML, from balancing conflicting business objectives to modernizing the end-to-end ML lifecycle using AI.

Here is what we have lined up for the night:

• GUAPO: Balancing Competing Interests Through ML Our very own Guilherme Baraúna and Yaroslav Marchuk (Data Scientists @ Glovo) will share how we tackle a massive internal challenge. Many teams at Glovo want to use the app's real estate to push for different objectives: loyalty targets subscriptions, growth aims at acquisition and retention, while ads push for immediate revenue. They will dive into how the GUAPO ML engine allows different stakeholders to compete for limited traffic using data-driven decisions, optimizing these tradeoffs in a unified way.
• Is AI Really Changing the ML Lifecycle? A Skyscanner Snapshot We’re thrilled to welcome Carles Riera (Data Scientist @ Skyscanner) as our external speaker! Carles will break down how Skyscanner's data science teams are currently using AI agents across the ML project lifecycle from ideation and scoping all the way to production and monitoring. He’ll cover the frameworks they use, real use cases in data analysis, model development, A/B testing, documentation, and also touch upon the areas where things haven’t actually changed that much yet.

🎤 AGENDA:

• 18:30 – 18:35 | Welcome & Glovo Presentation (5 min)
• 18:35 – 18:55 | GUAPO - Balancing competing interests through ML (Guilherme Baraúna & Yaroslav Marchuk) + 10 min Q&A
• 19:05 – 19:25 | Is AI Really Changing the ML Lifecycle? A Skyscanner Snapshot (Carles Riera) + 10 min Q&A
• 19:35 – 20:30 | Networking, Drinks & Food

⏰ WHEN Thursday, November 19th, 2026 | 18:30 to 20:30

📍 WHERE Glovo HQ (Yellow Park) - Carrer de Llull, 108, 08005 Barcelona

Carissimi, continuiamo il nostro viaggio itinerante per il #Veneto a #Vicenza !

Ci troviamo in #presenza giovedì 19 novembre alle ore 19:00.
La location sarà presso L'Elevator Innovation Hub di Vicenza. E per chi non potesse raggiungerci, si terrà anche in #streaming su YouTube.

La scaletta prevede uno speech seguito da discussione. 🗣 Mattia Carolo ci racconterà come fare girare un LLM su vLLM e infrastruttura Google GKE, attraverso le cose che si sono rotte davvero e nessuno tutorial ti ha mai detto.

E dopo, per chi vuole, resteremo lì per ( 🍸 | 🍺 ) e ( 🍕 | 🍔 )
e #networking ad oltranza ‼️

📽 E se avete piacere di condividere idee, progetti, argomenti, o proporvi come prossimi speaker, lascio un form Google e la pagina Sessionize !

----------- Please register below -----------

The next Basel Data Science & AI meetup is coming soon, this time with three highly practical presentations around various aspects of ML / AI deployment.
Join us for another evening of inspiration, networking and building the local data science and AI community together. Looking forward to seeing you there!

If you haven't joined our community newsletter, sign up here for monthly updates on events, jobs and more.

Agenda
18:30 - Arrival & mingling
19:00 - Talks (details below)
20:00 - Pizza, beers & networking

Talks

• Merve Selçuk Şimşek, Applied AI Scientist @ Institute for Data Science FHNW:
  How to Make Your Models More Efficient?
  As AI becomes increasingly widespread, the demand for computational resources and energy continues to rise, both during training and inference. Whether you're working with large or small models, generative AI or traditional machine learning, there are ways to improve efficiency, reduce costs, and minimise environmental impact.
  Merve holds a PhD in Computer Science and has spent the past decade working on making machine learning more efficient. In this talk, she will share practical approaches, real world examples, and the tradeoffs involved in improving model efficiency.
• Claudio von Planta-Thamboo, Staff Data & AI Engineer @ Swisscom: Turning Enterprise Knowledge into Action: AI Agents for Data Modernization
  How can AI accelerate data migration when the hardest work is understanding what already exists? Claudio von Planta-Thamboo will present his team’s approach to migrating a legacy data warehouse to a modern data platform on AWS. Their accelerator gives AI agents access to enterprise context through an ontology knowledge graph, helping them enrich metadata, align datasets with the established enterprise data model, and adapt legacy transformations and dashboards. Attendees will learn how knowledge graphs, human oversight, and reusable Model Context Protocol (MCP) tools fit together in this architecture, and how work done for migration can lay the foundation for AI-first data engineering.
• Boas Loeb, CPO & Co-Founder @ AIcoach LTD:
  Less is more: Why our AI coach learned to forget
  For an AI coach, remembering what users said before is crucial - but remembering everything turns out to be the wrong goal. Boas Loeb, co-founder and chief product officer of AIcoach.chat, an AI coaching platform used by organisations such as NHS Elect, shares how his team built an elaborate memory system and then had to simplify it. Today their coach’s memory borrows from how human memory works and is shaped by what coaching conversations actually needs: it keeps mainly what is worth exploring in a future session, links what keeps coming up again and again, and regularly forgets and reorganises the rest. Attendees will learn why more memory often makes AI less helpful, how to derive memory decisions from their own use case, and practical lessons for anyone building or using AI that “knows” its users.

The commitment fee of 5 CHF helps reduce no-shows goes directly towards elevating our apéro with some pizzas 🍕. If you have any issue with the payment method, let me know and we can figure something out!

Thanks a lot to Acuity Analytics for hosting the meetup and to Machine Learning Architects Basel for supporting us and providing the Apéro!

Speaker Tobias Pitters (German Center for Open Source AI & SHAP Core Maintainer)

Themen:
- Was ist Explainable AI?
- Wie wendet man explainable AI im Unternehmenskontext an?
- Wie schaut man in "Black Box" Machine Learning Modelle?

Hi Deep Learners,

We are happy to announce the next Vienna Deep Learning Meetup on November 25 at Fiskaly. Our Agenda:

***
Agenda:

• 18:15 Arrival
• 18:30 Welcome by the meetup organizers
• Introduction by the hosts
• 18:45 Talk 1: From Metadata to Labelled Imagery: MaskGIT for Land Cover Synthesis and ControlNet-Guided Diffusion for Earth Observation Images by Lydia Abady and Sebastian Ratzenböck (Another Earth)
• 19:30 Announcements
• Networking Break
• 20:00 Talk 2: Inside TED AI by Jakob Reiter (Mimirio)
• 20:30 Networking
• ~22:00 Wrap up & End

***

Talk Details:
--------------
Talk 1: From Metadata to Labelled Imagery: MaskGIT for Land Cover Synthesis and ControlNet-Guided Diffusion for Earth Observation Images
Pixel-level labels for Earth observation are expensive to produce and unevenly available across regions, classes, and resolutions. What if we could generate the labels first and then the images to match them? In this talk, we present a two-stage approach to generating synthetic image-label pairs. First, a MaskGIT-style masked token transformer synthesizes land cover maps conditioned on metadata such as class distribution, ecoregion, city, and ground sample distance. These conditions guide the composition and geographic characteristics of the generated scenes.
Second, a diffusion model generates corresponding optical imagery, using ControlNet to incorporate the land cover map alongside metadata conditioning. Because the image is generated from the map, every sample comes with labels aligned pixel by pixel.
Through a visual walkthrough of generated land cover maps and their matching images, we show how changing the metadata reshapes a scene, how the same layout looks across different regions and resolutions, and how this approach can produce labelled data on demand for rare classes and under-represented areas.

About the speakers:
Lydia Abady is an AI Engineer and researcher specializing in generative AI, machine learning, and computer vision. She holds a PhD from the University of Siena, Italy, where she researched the generation and detection of synthetic satellite imagery. Her experience spans academic research and industry, with a focus on developing innovative AI solutions for remote sensing and digital image forensics.

Sebastian Ratzenböck is an applied machine learning scientist developing generative models for Earth observation imagery. His work focuses on diffusion models, efficient generation, and scalable training and evaluation. He holds a PhD in Data Science and previously held postdoctoral positions at Harvard and the University of Vienna. His background spans generative AI, probabilistic machine learning, and simulation-based inference, with applications in astrophysics. He enjoys combining research with practical engineering to make models work with complex real-world data.

Talk 2: Inside TED AI
On October 28-30, 2026 TED AI is taking place in Vienna for the third time. It features high profile speakers such as Geoffrey Hinton, Peter Steinberger and a range of others. The talk topics range from AI & the brain, world models, drug discovery, security to AI & war.
Jakob Reiter is one of the main organizers of TED AI Vienna and will share insights from this year's edition of TED AI: a wrapup and summary of the main talks with key findings, as well as insights behind the scenes of "how to organize TED AI".

About the speaker:
Jakob Reiter is the CEO and Co-founder of Mimirio, a secure enterprise platform for knowledge and context management. Mimirio enables organizations to access trusted, AI-powered knowledge while keeping privacy at the forefront. As a tech entrepreneur with extensive technical experience, Jakob transforms complex concepts into practical, GDPR-compliant solutions.
As a member of the core team at TED AI Vienna, Jakob helps shape the conversation around responsible AI in business and society. He inspires leaders to approach digital transformation with clarity and responsibility.

We are looking forward to welcoming you at this meetup!
Please note that we have a max capacity of 85 persons at this meetup.
Your VDLM organizer team

Large Language Model (LLM) conference in London, United Kingdom

In a nutshell, extremal combinatorics studies how large a collection of finite

objects (numbers, vectors, sets, etc.) might be, if it has to satisfy given
restrictions. About a year ago when studying an extremal problem seeking a
six-tuple of different orderings that interlace in a rather peculiar way, we
have formulated and resolved a simplified variant of the original problem. In
May 2026, this simpler variant was added to the list of 100 problems in the
so-called Benchmarks in Leipzig, where state-of-the-art LLMs got tested on
research-level mathematics with known (human) solutions, and it remained as one
of the two problems that resisted the AI attempt.

In this talk, we give a light introduction to extremal combinatorics, and
describe the shattering problem. We will also briefly discuss math tools behind
our solution, emphasising the parts that were computer-aided (without any
AI/LLM tools). No mathematical background beyond high school will be needed.
-------------------------------------------------------------

⌚️ Start: 17:00 and end at 18:00, both online and offline in Truhlárna Karlín (Šaldova 388/5) - Onsite attendance is limited to 70 people.
🎙️ Speaker: Mgr. Jan Volec, Ph.D.
🍻 Networking after the seminar – great food and cold beer waiting for you!
🎥 Recording: After the event we will publish a recording and post a link to it in the comments.
🚪Doors open at 16:45, and the event officially starts at 17:00.
Your expertise is about to take off. Can't wait to have you on board!
-------------------------------------------------------------
Who is hosting the event
Miton is a Czech VC with portfolio companies like Rossum, Equilibre, or Rohlik. Apart from supporting Miton AI Times, Miton also issues the bi-weekly AI Newsletter.
Read more about Miton and AI.