Why it matters
Three practitioners from top-tier AI shops tackle real blockers: eval pipelines, session trace analysis, and enterprise deployment. If you're moving an agent from prototype to production, this is a rare chance to hear what's breaking in the wild.
The tokenmaxxing angle
Production agents burn tokens fast — optimizing via evaluations and session traces (the explicit agenda topics) is the primary lever for cutting inference spend. Understanding where agents loop or over-call tools directly translates to lower per-task token cost.
From the organizers
Three named speakers: Matt Nigh (GitHub, AI Program Manager Director), Ran Fu (Salesforce, Director of AI Agent Product Management), Raj Ganesh Jayaraman (AWS, Senior GenAI Solutions Architect). Runs 4:30–8:00 PM at Bellevue City Hall.