Production Agent Lab - Workshop

The infrastructure behind AI agents is changing quickly. New approaches are emerging faster than most teams can properly test them, making it hard to distinguish actual durable ideas from short-lived abstractions. We're doing a hands-on workshop day to shorten the gap between hearing about a new engineering approach and knowing whether it belongs in your stack. You’ll examine emerging trends to see how they behave in actual systems and develop a clearer sense of what to adapt or ignore. What to Expect Four technical, hands-on workshops Sessions led by teams building AI developer tools Practical demonstrations and real engineering workflows Opportunities to build, test and debug alongside other engineers Live demos to close the day Workshop lineup Build a Competitive Intelligence Agent with Live Web Data With Amaury Desrosiers, Head of FDE and Director of Solution Consulting @ Nimble Learn how to connect your AI agent to the live web using Nimble, then put it to work building a competitive intelligence workflow. We’ll walk through gathering current competitor information and turning it into a clear, sourced briefing. You’ll learn an approach you can adapt to your own research and monitoring use cases. Loop Engineering: Design and Observe the Loops Running Your Agents With Abby Morgan, AI Research Engineer and Developer Advocate @ Comet Move beyond brittle, predefined paths by designing loops that let agents act, inspect results, recover from errors and decide when to stop. You’ll define reliable exit conditions, distinguish recoverable failures from hard blockers and instrument the loop with Opik to trace each iteration and measure whether your changes actually improve performance. More workshops will be announced soon. Agenda 10:00 AM — Doors, coffee and check-in 10:20 AM — Keynote 10:30 AM — Workshop 1 11:30 AM — Workshop 2 12:30 PM — Lunch 1:30 PM — Workshop 3 2:30 PM — Workshop 4 3:30 PM — Demos and closing 4:00 PM — Event ends Note: This event is fully in-person in San Francisco. Capacity is strictly limited. All attendees must apply and be approved prior to the event. Unapproved or pending applicants will not be admitted at the door. Who should attend? Mid-to-senior engineers, technical leads and hands-on founders building or exploring AI agents. Familiarity with LLMs, APIs and software engineering is recommended. Our Lab Partners Comet · Open-Source LLM Observability and Evaluation Nimble · Real-time web data infrastructure for AI Nexla · Enterprise-grade Data Platform For Agents See what’s possible. Follow us on X · LinkedIn · Newsletter