02 / IT · Analysis
QCon AI New York programme covers agents and code review
Engineers’ accounts of running AI systems frame a programme covering agent permissions, code review and team adoption.

InfoQ reported the publication of QCon AI New York’s programme of engineering talks on AI systems used in development and operations.
According to InfoQ, the programme centres on practising engineers’ accounts of running those systems. Agent permissions and AI code review place the use of coding tools within the broader work of operating AI in engineering teams.
InfoQ also identified team adoption as a programme topic. The scope connects the systems engineers operate with the ways development teams put AI tools to use.
Engineers presenting systems they operate
InfoQ reported that engineers from LinkedIn, Netflix and DoorDash are among the speakers presenting AI systems they run in production. Their talks centre on systems in operational use.
According to InfoQ, engineers from Figma and OpenAI also feature in the programme. The speakers’ own systems provide the basis for examining how engineering organisations use AI.
Agent permissions, evaluation and operating costs
InfoQ listed agent permissions and guardrails among the talk topics. The programme also covers evaluation loops, the repeated assessment of AI outputs. These topics put controls on agents alongside checks on their results.
InfoQ also identified inference cost—the cost of generating model responses—as a topic. Alongside code review and team adoption, it gives the programme a common focus: the work involved in operating AI within engineering organisations.