Agents in Production: What Breaks, What Works, What You Should Know

Overview

Agents fail. The useful part is figuring out why, catching it sooner, and making sure the next agent doesn’t repeat the same mistake. Join fellow builders to compare what’s working, see what others are building and share lessons from running agents in production.

Who it’s for

Software engineers, AI/ML engineers, platform and infrastructure teams, and technical founders building, deploying, or debugging AI agents. Whether you’re moving a prototype into production or already running agents with real users, come compare notes with other builders.

Agenda

5:30 – 6:00 PM - Check-in, food & networking
6:00 – 6:40 PM - Panel Session
6:40 – 7:00 PM - Builder demos: 4 attendee teams, 5 minutes each
7:00 – 8:00 PM - Networking & food

About the hosts

Oodle.ai - Observability built for agents. AI can't debug what it can't see. With Oodle, agents debug applications, infrastructure, and other agents with full context. Teams use it to build self-improving agents and catch failures early. One platform for full-stack and agent observability, built on object storage and serverless compute so you can store everything, query anything, and run sub-second agentic queries at 80% lower cost.

XTrace - Enable engineering teams to ship reliable software faster by turning lessons from individual coding sessions into shared knowledge and enforceable coding standards. It puts what one engineer learns to work across the team’s AI coding agents, helping prevent repeat mistakes, reduce costly rework, and get new features to customers sooner.

Infron AI - Production inference infrastructure for AI applications and agents. Infron gives engineering teams one OpenAI-compatible API for 400+ models from 100+ providers, with routing and fallback through the same integration. Teams can test and switch models without rebuilding for each vendor, then route workloads across providers as reliability and cost requirements change. Eligible models are available at up to 35% off direct pricing, helping teams scale production inference with fewer infrastructure tradeoffs.