The conference, scheduled for December 15–16 at The Westin Jersey City Newport, targets senior engineers, architects, and technical leaders building AI systems in production. Conference Chair Hien Luu noted that the program signals a broader shift: "AI engineering has become systems engineering." According to Luu, the challenge is moving from model behavior to system behavior, requiring agents with bounded execution authority, context management, and probabilistic models wrapped in deterministic control planes. Topics include harness engineering, continuous evaluation, observability, and policy enforcement as core infrastructure.
Keynote speaker Nancy Wang, CTO at 1Password, will explore identity and authorization for agents in production in a talk titled "When Software Becomes a User: Identity and Authorization for Agents in Production." Wang will examine how traditional identity models—designed for single human principals with stable roles and bounded sessions—break down when agents act for multiple users, invoke unforeseen tools, create subagents, and operate without supervision. The session will cover delegated authority across chains of agents, auditability for multi-hop tool calls, and techniques to grant agents sufficient access without exposing credentials.
Other confirmed sessions include a talk on guardrails for operations agents running on large-scale Kubernetes infrastructure by Ronak Nathani, as well as discussions on shared model-serving platforms and post-deployment evaluation of AI decision systems. The full program is available on the QCon AI New York website.