面向自主云MLOps的前沿全栈工程框架
原标题：Forward-Deployed Full-Stack Engineering for Autonomous Cloud MLOps
AI 摘要
本文提出了一种基于证据门控的多智能体框架，用于将自然语言描述的MLOps云工程任务转化为经过验证的代码仓库和可运行的云部署。该框架结合了图工程、循环工程和智能体工程，通过有状态的图编排器协调仓库生成、审查、执行、验证、发布和监控等专业智能体，并仅在可验证的执行或运行时证据支持时允许关键生命周期转换。研究在Google Cloud Platform上实现了该框架，实验结果表明其能有效防止不支持的生命周期转换，并引导每次运行达到已验证的部署或可审计的终止失败。
正文节选
Forward-Deployed Full-Stack Engineering for Autonomous Cloud MLOps Abstract Across industries, machine-learning systems support applications ranging from prediction and anomaly detection to forecasting, optimization, and scheduling, yet operationalizing these systems requires coordinating application development, model pipelines, cloud infrastructure, security, deployment, monitoring, retraining, recovery, and rollback. We present an evidence-gated multi-agent framework for transforming a natura