Four talks across three perspectives, one evening.
From enterprise — on agents in real production: what shipped, what broke, what it cost, and what had to change in the organization around it. From research — on what the evidence actually says: evaluation, failure modes, human–agent interaction, the science under the tooling. And from the edge — on the hot topics we're all improvising through in public.
Intended for everyone using Claude Code and agentic tooling regularly, who wants to share and learn from the community. Especially on regulation, now that the EU AI Act is no longer someone else's problem. On audit trails and code review when much of the code wasn't typed by a human. On hyperscalers, data residency, and the procurement conversation nobody enjoys. On sandboxing, or how much rope to give an agent. And on the jump from one agent to many.
Plus time to talk to each other.
Tentative plan:
17:00 Welcome17:10 Talk 117:30 Talk 218:00 Food & Drinks18:30 Talk 319:00 Talk 419:30 Networking20:00 End
Talks are 20 minutes plus questions. Programme is tentative until speakers are confirmed.
Speakers
Announced here as they're confirmed. Four slots: one or two from enterprise, one from research, one from the edge.
Want to give a talk?
We have open speaker slots. We're looking for people who have done the thing, and we'd rather hear one honest account of something that went wrong than five that went well.
Especially interested in going from a single agent to a team of them and what breaks along the way: orchestration, delegation, merge chaos, mapping your existing workflows onto agents and discovering which parts refuse to map. And the road from "worked on my laptop" to something a team depends on — evals, cost, observability, rolling it out to N people and proving it paid off.
Hosted by Diana Meda, Claude Community Ambassador, as part of the global Claude Community Events programme supported by Anthropic and by Ahcène Boubekki, postdoctoral researcher at the Statistics and Probability section of the Mathematics department of the University of Copenhagen.
This edition is made possible with the great support of the the Danish Data Science Community (DDSC) and IDA, who are helping with venue and with reaching their communities.
