智能体评估训练营：6月27日5小时线上实操工作坊，聚焦AI智能体评估
英文摘要
A live, hands-on bootcamp on evaluating AI agents will be held on June 27, led by AI engineer Ammar Mohanna, PhD. The 5-hour session covers four evaluation layers: component, trajectory, outcome, and adversarial evaluation. Attendees receive a practical evaluation framework, 6 months’ access to an AI Evals assistant, implementation templates, a capstone project, and a Packt-endorsed certification. The event targets teams that struggle with agent failures in production due to poor evaluation practices.
中文摘要
一场由AI工程师Ammar Mohanna博士指导的线上实操训练营将于6月27日举行，时长5小时，覆盖组件、轨迹、结果和对抗性四个评估层面。参与者将获得可立即应用的评估框架、6个月的AI评估助手使用权、实操模板、一个结业项目以及Packt认可的证书。活动旨在帮助因评估不足而导致智能体在生产中失败的技术团队。
关键要点
Date and format: June 27, 5-hour live online bootcamp with hands-on exercises.
时间与形式：6月27日，5小时线上直播，包含实操练习。
Instructor: Ammar Mohanna, PhD, an AI engineer and researcher specializing in production AI agent evaluation.
主讲人：Ammar Mohanna博士，专攻生产环境智能体评估的AI工程师与研究员。
Content: Evaluation framework covering component, trajectory, outcome, and adversarial layers.
内容：涵盖组件、轨迹、结果和对抗性四个层面的评估框架。
Deliverables: 6-month AI Evals assistant access, implementation templates, a completed capstone project, and a Packt-endorsed certification.
收获：6个月的AI评估助手使用权、实操模板、完成一个结业项目，以及Packt认可的证书。