Selected from 344 submissions from 38 countries: on 20 November 2026, the Marriott Zurich brings together engineers who don't just talk about agents but build them – and let them build.
The Agentic Engineering Days Zurich 2026 are Switzerland's conference for AI coding and agents. One day, four tracks, 40+ sessions and 50+ speakers – on 20 November 2026 at the Marriott Hotel next to Zurich main station. Every contribution was hand-picked from 344 submissions from 38 countries, against one clear criterion: technical depth and honest experience reports, including what didn't work.
Keynote and international highlights
The keynote is delivered by Andrés Marafioti, Multimodal Research Lead at Hugging Face, with Reachy mini: Giving a body to AI. Ricardo Silveira Cabral (NVIDIA, ex-Meta FAIR) presents Something New from NVIDIA, and Copilot co-creator Krzysztof Cieślak of GitHub Next looks ahead in The next frontier – agentic collaborative environments.
Swiss practice: what actually runs in production
The programme is particularly rich in experience reports from Swiss companies. Migros Online shows in Agentic AI at Migros Online: Autonomous Engineering Agents and Conversational Shopping how autonomous engineering agents and conversational shopping work together day to day. Swisscom reports in Beyond Prompts & Models: Evolving Enterprise Agentic Systems on how agentic systems are evolving inside the group. Swiss Post makes the case for small language models over large ones in OCR at Swiss Post: The Case for Small Language Models.
Add to that Securing MCP in the Enterprise: OAuth, Token Exchange, and Real-World CI/CD Integration by ELCA, A Practical Guide to Fast and Private LLMs by bespinian, You Can't Ctrl-Z a Robot: Coding Agents in Embedded Systems by adesso Schweiz, 21 million exceptions a day: getting Bedrock and AgentCore right by BLP Digital and copebit, as well as From a RAG Demo to a RAG System by amnis.
Track 1 "Let it build": building software with agents
Databricks shows how everyday development changes when several coding agents work in parallel in Human-in-the-Loop Is You: A Guide to Working With Multiple Coding Agents. Anyone who wants to know what happens under the hood should go to Let's build our own Claude Code. Measurability is the theme at Xebia with We Rolled Out GitHub Copilot… How Do We Prove It Helps? and at Flagsmith with Tie a Metric to Every Release.
Things get controversial in AI & Agility: Is Spec-Driven Development Bringing Back Waterfall? – complemented by the counter-position from practice in Specs first, code later by AWS. Also in this track: Migration or migraine: What no one tells you about using LLMs, Close the Lid: Coding with AI Agents from Your Phone, From Copilot to the Dark Factory, The Agentic Lab Factory: How I Stopped Writing Labs and Started Reviewing Them and When Software Stops Being Deterministic: Rethinking Engineering Practices for AI-Native Systems by Typeform.
Track 2 "Build it": agents in production, security and governance
Anyone running agents knows the uncomfortable questions. Wrike tells those stories in "YOU MUST NOT ssh!" and Other Stories from Building Our SRE Agent. Alongside it: Watching Agents for Trouble on observability, Running Secure Agents at Scale by Adobe, Building Authorization for Agentic RAG Systems by AuthZed and Agents Speak Protocol: Why Standards Are the Real Infrastructure of AI by IBM.
Relevant for decision-makers: Deploying Agents at Scale: What Enterprise Adoption Actually Looks Like by Accenture, FSI Transformation at Scale: Sovereignty & Compliance by Atlassian, Your Biggest New Customer Can't See Pixels by ABN AMRO and Caveman and the Empty Room: Designing Agents for an Audience of Zero.
Hands-on and panel
In the hands-on sessions you build instead of listen, 75 to 90 minutes each on your own laptop: Just Build — Orchestrating your Agents with Codex with OpenAI, ADK Crash Course – From Beginner To Expert with Google Cloud, Build Your Own Production-Ready Agentic Apps with IBM and Red Hat, Hands-on Spec-Driven Development as well as Agent on the Wire: Looking Inside an AI Coding Agent.
The tracks close with the panel discussion Can We Trust Autonomous Agents? featuring representatives from Microsoft, Swiss Life, TheRedCode and VisionPilot, moderated by Maxim Salnikov.
Practical details
The schedule for the day is set: registration from 07:30, opening and keynote at 08:30, talks and hands-on sessions until 17:00, followed by apéro and networking until 18:45. Exact times and rooms will be published at the end of October – all sessions and all speakers are already online.
Regular tickets cost CHF 442, with a 20 percent discount from three tickets onwards, and companies can pay by invoice instead of credit card.
View the programme and get your ticket →
Agentic Engineering Days is organised by Letsboot.