ETH AI Center Researchers Community Meetup
Recurring meetup series for the AI research community at ETH Zurich.
59 upcoming meetups across Switzerland, checked daily.
Recurring meetup series for the AI research community at ETH Zurich.
Two talks on what AI has actually changed in building products: Coople on a year spent replacing squads, sprints and story points with AI-first workflows and smaller teams, including the experiments that failed, and Brack.Alltron on where AI earned its place and why plain analytics sometimes still beats machine learning.
Monthly reading group in Bern where the group works through one applied-AI article together, with time afterwards for open exchange; the paper is shared in advance.
A beginner-friendly hands-on evening building a working web app from scratch with Kiro, AWS's agentic IDE, with the organisers on hand to help — bring a laptop, pizza and apéro included.
Three talks and an apéro: the guardrails Open Systems put in place before letting an AI agent work on its Kubernetes clusters, a Rust-based zero-dependency etcd replacement with the QuantyKube sandbox, and a session on Jellyfin.
The group’s last evening of 2026: Hidora on building a highly available multi-availability-zone cloud in Switzerland on Kubernetes, and a talk on defining quality processes across several teams in scrum-of-scrums delivery.
Dr Stamen Popov on where AI already adds value from drug discovery to patient support, and where trust, ethics, regulation and adoption still get in the way of better outcomes.
Launch evening for Giotto Alto and Solo: meet the team, try the models, watch a Giotto Workstation demo and argue about what sovereignty in AI actually requires — with pizza, drinks and live music.
V. Moskvoretskii (EPFL) on building an assistant persona into pretraining rather than bolting alignment on afterwards — a technique now part of training Apertus 2, Switzerland’s open model. Moderate technical background assumed.
A German-language lunch hour on the organisational side of Microsoft 365 Copilot rather than prompt tricks: Roman Schön on why AI initiatives stall, why structure and context beat perfectly worded prompts, and what it takes to make the tool stick in day-to-day office work.
First evening of Global AI Zürich’s new agentic series, with three talks on real agent implementations covering identity, software factories and agent memory.
An evening with no fixed programme in place of the usual weekly session: meet the organisers, ask how the AI safety field is put together, or pick up a thread from an earlier talk — newcomers as welcome as people already working on it.
Open evening where data science students present their end-to-end projects, from data collection and analysis through modelling to visualisation, followed by questions and networking.
Kick-off meetup for a new Basel group around AI-assisted coding: short intros, side-project show-and-tell, and an open round on what people want to build next.
A relaxed Saturday-morning coworking session where builders and AI-curious people work on their own projects and swap prompts, workflows and lessons learned.
A demo-first evening for people actually building with AI: around half a dozen five-minute demos of work in progress, each followed by questions — slides and pitch decks are explicitly not allowed — then pizza and networking.
Three practical talks on getting machine learning and AI models deployed and running in production, followed by networking with the Basel data community.
Four speakers from Google, AWS, Unique and Systematic-X on keeping AI systems trustworthy once they act on their own — how to evaluate and monitor before and after deployment, where guardrails, human oversight and red-teaming fit, and how to ship fast without losing accountability.
Lightning talks from the interpretability and safety hub built around Apertus, Switzerland’s open multilingual model: shared representations across languages, emergent misalignment from stereotypes that pass moderation, probes that catch repetition loops early, and persona binding formed in pre-training.
A map of the routes into AI safety and governance work — fellowships, research programmes, who is eligible, when they open and how they are funded — with people who have been through them answering questions. No particular background needed.
Designers and developers from Interactive Things walk through how they turned qualitative peace indicators from around the world into a comparable visual dataset.
Kouros, who is not a developer and has a full-time job, explains how he used AI to build a custom ERP for the La Nébuleuse brewery, followed by guided discussion, audience questions and an apéro.
Monthly Geneva AI evening with three practical talks — building with LLMs, working faster with AI and where it is all heading — then pizza and networking. Forty places.
Breeze Dong on the operating model behind her book The Living Organization, and an open discussion of what happens to teams, roles and decision-making when AI agents join the work.
A French-language Azure evening in Geneva: what AI has really changed about Microsoft recruitment and how candidates stay visible, then Azure Local for running Azure services inside a sovereign private cloud, with a live Foundry Local demo — quiz and networking afterwards.
Two talks and an apéro hosted by Puzzle: wiring Amazon RDS to TimescaleDB with foreign data wrappers for federated Postgres on EKS — and where 10 TB starts to break it — then PostFinance on moving cloud and AI cost awareness earlier, and what AI cost governance takes.
Short talks and paper walkthroughs on learned world models for robotics — action-conditioned video prediction, joint observation and action models, planning and policy evaluation, and synthetic data generation.
Evening session on how far AI agents should be allowed to act on our behalf, with Maria Fay (SAP, University of Münster) and Matthias Plattner (Julius Bär) on AI governance across industries and whose values an agent should carry.
Stefan Gilgen runs a live, unrehearsed experiment in which several AI models write the text, generate the images and assemble a complete website on stage, with a candid look at which models he uses day to day and what breaks along the way.
Fourth session of the AI Futures series, turning from the risks to the responses: technical safety work such as alignment, control, interpretability and evaluations; governance and policy including compute rules and treaties; and socio-structural preparedness — with the organisers’ own read on what looks most promising.
Swiss MVPs recap the announcements from Microsoft Ignite 2026 in short sessions on Cloud & AI, the agent platform, Copilot and Microsoft IQ, security, data security and compliance, and agentic business applications, with an apéro after.
An in-person working session where the group picks one AI agent idea from the room and explores it together with sticky notes, sketches and a timer — open to beginners and active builders alike.
An Italian- and English-language evening in Ticino on how recent, well-known cyberattacks actually worked and how generative AI is shifting the balance between attackers and defenders, with a social-engineering quiz and a networking aperitif.
In-person evening at Digicomp on the infrastructure side of running AI workloads on Azure.
Final Bern meetup of the year for the Swiss AWS community: a mix of community talks plus discussion and networking, with the agenda announced closer to the date.
The closing session of IBM's 2026 hybrid series on IBM Bob, its AI development assistant — coding live alongside other European cities, with beginner, intermediate and expert tracks covering everything from first idea to modernising legacy applications.
Build-and-deploy evening at Nine: nxt Engineering on how not to wreck an Azure migration, and Access for all on why manual accessibility testing still matters alongside automated and AI-assisted checks. Also livestreamed.
A German-language breakfast session with Quatico CEO Jan Wloka on why adding more AI agents tends to produce more review work than finished work, and the three changes that fixed it for his team: a written decision before any agent starts coding, work split into packages that run in parallel, and one daily overview of what needs a human. Aimed at product owners and delivery leads, with a real work package run live.
Giacomo Castiglioni of Huawei’s Computing Systems Lab on compiling for asynchronous execution units on Ascend NPUs, with a scheme that lets the compiler insert and verify synchronisation for Gated DeltaNet and FlashAttention kernels.
Free German-language evening in Gebenstorf: a current AI update from Michael Kyburz, then Alexander Repenning on what AI means for children, closing with an apéro with the swissAI community.
Show-and-tell evening for people building robots: short demos of works in progress, honest accounts of what broke, then food and conversation.
A live look at an agentic setup that has been in daily office use for months: the agent prepares a project kick-off from a single instruction, turns a correction into a standing rule, and flags an email carrying a hidden instruction.
The season finale of Google Cloud's Founder's Story series: a panel on whether Europe can build its own global AI companies or will mostly buy AI from elsewhere, covering research, open models, compute, sovereignty and capital, followed by a networking apéro at Google's Europaallee office.
Around six five-minute live demos of AI tools, platforms and works in progress — no slides allowed — with a legal-tech slant for this edition, followed by networking and an apéro.
Evening for the Zürich dbt community aimed at analytics engineers, data engineers and data leaders, opening with Massimo Turola of Bank Vontobel on using dbt to manage Snowflake Cortex AI applications in an AI data catalog, then an apéro.
Arnaud Ricci on how he built Genève 3D, a browser-based 3D model of the whole canton, using AI agents and open data — with a live demonstration, questions from the room and a non-alcoholic apéro. The talk is in French with English subtitles.
An evening for database practitioners hosted by Swissquote: live demos and case studies on database modernisation, high availability with MaxScale, an Oracle-to-Galera migration, and pairing real-time payment workflows with in-memory caching and vector search for AI.
An interactive session on where enterprises genuinely need autonomous AI agents and where a deterministic AI workflow is the better answer, working through real-world cases and the trade-off between autonomy and control.
The Rhine valley edition of OST’s monthly AI round table: an informal evening in the Buchs CREALAB where university experts talk through how the technology actually works and how to point it at a problem you already have.
Two talks in Zurich: Red Hat's Simon Krenger on running virtual machines next to containers with KubeVirt, including live demos and when it is the wrong tool, and Basia Karaagac on where AI genuinely helps in QA and where exploratory testing and human risk judgement still win.
A panel with Peter Fankhauser (co-founder and CEO of ANYbotics) and Anne Scherer (co-founder of Delta Labs) on how AI is changing company operations, from autonomous robots on the plant floor to generative AI in the boardroom.
Four practitioner talks for CTOs and engineering leaders on AI in software delivery: a year of lessons from building with AI, securing the software lifecycle against prompt injection and compromised agents, making enterprise data ready for AI agents, and AI-assisted delivery inside regulated healthcare.
Closing session of the AI Futures series, moving from grand plans to personal ones: how people actually get into AI safety work, what to do about personal finances under uncertainty, and how to handle information overload and keep your footing.
Prof. Stephan Mandt (UC Irvine) on generating several mutually consistent tokens in one forward pass — models pretrained to 3 billion parameters that approach autoregressive quality with inference up to 5.7× faster. Hosted by Prof. Julia Vogt.
Data-ethics community meeting at the UZH Digital Society Initiative: data-quality lessons from the City of Zurich's MeLODy chatbot and the Federal Statistical Office, a study on human-AI teaming in marketing, and the “AI Encyclical” dialogue format, followed by an apéro.
The Basel AI Stammtisch returns for the new year with a guest talk or an in-house topic, followed by open discussion over drinks at the iRIX bistro.
The Swiss cloud-platform community spends an evening on one question: what changes when AI agents, rather than people, become the most important users of an internal platform.
Hybrid Databricks user-group evening in Zurich on automating data work with AI-enabled workflows, including a session on Databricks Apps and Lakebase applied to clinical cancer research cohorts. Rescheduled from November 2026 to February 2027.
Tenth edition of the Graubünden AI meetup in Chur, with Rico Castelberg on voice AI and digital telephone assistants, followed by open discussion and networking.
No events match those filters.