Copenhagen | Claude Code for Developers

Welcome to the ninth instalment of the Copenhagen Advanced Machine Learning Study Group (CAMeLS-Group). We meet every second Monday to read and discuss a machine learning paper together at the "Snaren" reading area on the fourth floor of the Copenhagen Municipality Central Library. In a time of endless online chatter and AI-induced attention decay, we’re trying something entirely old: reading and discussing together in person. The idea is simple: bring your laptop or a printed copy of the paper - BUT DO NOT READ IT IN ADVANCE! During the session, we read each section individually at the same time, then pause to discuss it together before moving on. Once we’ve finished the paper, we come together for a broader discussion of the paper as a whole. Recommended prerequisites: some background in ML or related fields, at a level where you can read a paper in your favourite area. Don’t feel scared off if you have not yet graduated, but we will be reading research papers. In this session, we will be reading the paper: WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution. The paper can be found here: https://arxiv.org/abs/2608.27454 If you want to participate in the organisation of future events, please join the Danish Data Science Community Slack workspace and go to the CAMeLS-Group channel: https://join.slack.com/t/danskdatascie-o8m9638/shared_invite/zt-4975uw282-6M1vlRA50ck4uAg75CrluA

Claude Impact Lab — turning questions into working ideas Join a Claude Community Event, organized by local enthusiasts for everyone who loves building with Claude! A hands-on afternoon and evening exploring one central question: How is AI changing the way we work—and what should we build in response? This Impact Lab builds on the conversations and questions raised at Claude Conversations, back in August. Instead of stopping at discussion, we will turn the most important themes into practical experiments, prototypes, and concrete proposals. The format combines short inspiration, guided challenges, collaborative building, and a final demo session. Participants will work in small groups around a set of focused questions, using Claude and other AI tools to explore possible solutions. The goal is not polished products; it is to learn quickly, test ideas, and leave with something tangible. What to expect Short opening talks and an introduction to the Impact Lab challenges A structured, hands-on build session in small groups Guidance and support throughout the afternoon Informal food, snacks, and time to meet other participants Claude API credits to build and original merchandise from Anthropic for participants A closing demo hour when teams share what they explored or built From ideas to impact The Impact Lab is designed to be a starting point—not the end of the work. Every team will define the change it wants to create and leave with something that can be taken forward: a prototype or concrete concept, the people it is intended to help, and a clear next step. Alongside their solution, teams will create a simple 30-day impact plan with a named owner, the first real-world test, and the support needed to continue. During the demo hour, teams will share both what they built and what they will do next. After the event, the projects and commitments will be documented and shared with participants. Teams that want to continue can stay connected, invite new contributors, and repo

Your AI bill is the fastest-growing line in your P&L, and most teams can't say what one request costs them. Agents make it worse: one user action becomes 3 to 30 model calls, each with its own context, tool calls, retries and reasoning tokens. The curve isn't linear, and credits only hide it for a while. This is a hands-on evening for Pre-Seed, Seed and Series A founders and technical employees who are building with LLMs today. No product slides. Three people who see hundreds of startup AI bills will tell you where the money actually goes and what the teams that fix it do differently. Then we split into small rooms so you can bring your own numbers. Speakers Philip Andersson, Applied AI @ OpenAI Andy Kane, Specialist, GenAI Models @ AWS Andree Hultgren, AI Engineer @ Lovable Moderated by Deina Kellezi, Investor at byFounders. What we'll cover 1. Know your cost per outcome Cost per task as a first-class engineering metric. Tag spend by feature and tenant, set budgets and alerts, attribute per key and service. Catch the runaway test script and the leaked API key before the invoice does. 2. Cut tokens without cutting quality The levers in effort order: model right-sizing, cascading and routing, prompt caching and batch, context trimming and output caps, prompt compression, fine-tuning or distillation. And the rule that makes them safe: no cost change ships without an eval. 3. Agents: where the bill goes non-linear Per-run budgets and step limits, circuit breakers on loops, tracing cost per trajectory rather than per call, managing context across steps, and evaluating on outcome plus cost. Closing with the hard question: how do you price something whose cost varies 10× per task? Agenda 16:30 Doors open, drinks 17:00 Panel (30 min) 17:30 Breakouts: each speaker hosts a room on their topic, you choose the room. Bring your questions. 18:15 Walking dinner and drinks 21:00 Close Who should come You're a founder, CTO or early engineer at a pre-seed or seed company, you hav

We're inviting PostHog to the stage to talk about their newly launched Replay Vision, which watches your session recordings for you - so YOU can go birdwatching 🕊️👀 PostHog has been automating for self-driving, and they will share what Replay Vision caught and how it helps build successful products. Then the floor is yours: Demo your products and how you've used PostHog to improve them. Schedule: 6:00 - 6:30 pm: Check-in, arrivals, food 6:30 - 6:40 pm: Birding for beginners, seriously 6:40 - 7:00 pm: PostHog on Replay Vision 7:00 - 7:30 pm: Community Demos 7:30 - 9:00 pm: Networking and casual chats Limited spots. Demos, food, merch, and binoculars provided.

Breakfast with Heyra and ElevenLabs on the current state of voice agents, what they can handle today, and how companies in regulated industries are starting to use them. Most people associate voice bots with rigid menus, keyword recognition and repeating their CPR number over and over. That has changed. Voice agents now hold real conversations, handle sensitive data and work around the clock, and companies in pensions, finance and insurance are starting to use them. Join Heyra and ElevenLabs for breakfast and a practical look at where the technology stands today and where it fits in your organisation. What we will cover: A live demo of what voice agents can do today, and a quick look back at how far they have come. Voice agents in customer service, with Sæmundur Ragnarsson, AI Engineer at Heyra: how today's conversational agents differ from dialogue-tree bots, and how integration with your own data and systems builds trust. Use cases beyond the call centre, with Nils Smed, Partner at Heyra: voice agents we have built for work outside customer service, and how to judge which conversations in your organisation an agent should take on. A fireside chat and Q&A with Anna Neely and Emma Hernmarck from ElevenLabs, hosted by Nils Smed. Anna leads ElevenLabs' forward deployed engineering team for banking, financial services and insurance in Europe and works directly with its most regulated customers. The conversation covers how enterprises in regulated industries handle trust, security, risk and AI disclosure, how to scope a use case for a demanding deployment, and how voice goes from a first use case to a core channel. This session is for leaders who are responsible for, or closely involved in, how their organisation serves its customers, from customer service and operations to digital, technology and compliance. Host: Heyra, together with ElevenLabs Speakers: Sæmundur Ragnarsson and Nils Smed (Heyra), Anna Neely and Emma Hernmarck (ElevenLabs) Format: Breakfast, talks, l