Zürich AI | Summer Meetup

Join us for Summer Zürich AI Meetup

Agenda:

– AI as Your Copilot – Supercharging Product and UX by Omar Moling (Director of Product) & Daphne van de Brule (Director of Product Experience) from Beekeeper
– Building Production-Ready AI Systems: Lessons from UniQlearn by Sam Naji from UniQlearn
– Agent-to-Agent (A2A): Standardizing AI Communication by Alex Erfurt from Google

OPEN MIC SESSIONS 🎤
For the first time, we will be hosting Open Mic 🎤 sessions, giving everyone an opportunity to present recent work in AI/ML. If you’ve been working on something interesting lately, come and reserve a 5-10 minute slot to showcase your project to everyone

Abstract:

AI as Your Copilot – Supercharging Product and UX
AI is reshaping how we build products, structure teams, and scale organizations. In this talk, we’ll explore how to develop an AI strategy that evolves from assistant features to agentic experiences, agent-to-agent collaboration, and even AI agent builders. We’ll also discuss how product organizations can adapt their structures, roles, and ways of working to stay ahead in this new landscape. This session offers a practical example of how AI is being applied at Beekeeper.
---
Building Production-Ready AI Systems: Lessons from UniQlearn
In this talk, I’ll share practical insights from leading AI development at UniQlearn, where we built scalable, kid-safe, multi-agent LLM systems used by thousands. I’ll walk through the real-world practices we use to take AI products from prototype to production, including prompt engineering workflows, continuous feedback loops with Langfuse, safe model deployments with LangGraph. This session will be valuable for teams looking to build reliable, observable, and evolvable AI systems that go beyond demos and into daily use.

---
Agent-to-Agent (A2A): Standardizing AI Communication
The A2A (Agent-to-Agent) protocol establishes a standardized framework for AI agents to communicate and collaborate. Its core aim is to enable agents to discover each other's capabilities, negotiate interaction modalities, and securely execute tasks cooperatively. Among diverse multi-agent systems, this facilitates seamless information exchange and joint task completion.