OneAdvanced如何在英国主权AWS上部署超过50个AI代理
摘要
本文介绍OneAdvanced如何在英国主权AWS上通过自托管Llama 4 Maverick和Llama Guard 4,构建包含50多个专用AI代理的UK-sovereign AI解决方案。
核心要点
- OneAdvanced服务超过10,000家客户,主要面向医疗、法律等受监管行业,对数据主权要求严格。
- 由于Llama 4 Maverick和Llama Guard 4当时未在英国区域托管,OneAdvanced选择在自有AWS账户中自托管开放权重模型,部署在伦敦(eu-west-2)区域。
- 模型通过vLLM部署在Amazon SageMaker AI上,使用p5.48xlarge实例,并针对120K-128K token上下文窗口进行优化。
- 超过50个Strands代理运行在Amazon ECS上,每个代理有独立的系统提示和工具配置,配置存储在DynamoDB中。
- RAG流水线使用Amazon Aurora PostgreSQL与pgvector扩展,文档通过S3上传后转换为markdown、分块并嵌入。
原文佐证
- OneAdvanced, a UK-based enterprise software provider serving over 10,000 customers, needed to deliver AI capabilities while making sure that no data would leave the UK.
- At the time, the specific models they wanted, Llama 4 Maverick and Llama Guard 4, weren’t yet available through managed services in the UK region.
- This architecture enabled OneAdvanced to achieve full UK data sovereignty and rapid agent deployment, supporting their ISO 42001 certification for AI governance while maintaining complete control over their model serving infrastructure.
AI 洞察
这一案例表明,在主权云环境下,自托管开放权重模型正成为企业满足数据合规需求的关键路径。随着Llama类模型在英区可获取性提升,企业可逐步从自托管转向托管服务,但自托管带来的控制力仍具吸引力。对于AI基础设施提供商而言,支持主权云上的自托管部署将成为差异化竞争力。