OneAdvanced deploys 50+ AI agents on UK-sovereign AWS
50 specialized agents run entirely within UK data centers using Llama 4 Maverick and Llama Guard 4
OneAdvanced, a UK-based enterprise software provider, deployed over 50 AI agents on a UK-sovereign AWS architecture to meet strict data residency requirements for its 10,000+ customers in regulated industries. Faced with the challenge of model hosting and data sovereignty, the company self-hosted Meta's Llama 4 Maverick and Llama Guard 4 on Amazon SageMaker AI using p5.48xlarge instances in the London (eu-west-2) region.
The solution integrates a RAG pipeline with Amazon Aurora PostgreSQL-Compatible Edition (using pgvector for embeddings), over 50 specialized agents powered by Strands Agents SDK, and a tool layer running on Amazon Elastic Container Service (Amazon ECS). Llama Guard 4 filters harmful inputs before they reach the main model, while agents retrieve documents or invoke specialist tools like web search. This architecture enables OneAdvanced to achieve ISO 42001 certification for AI governance while maintaining complete control over its infrastructure.
- Self-hosted Llama 4 Maverick and Llama Guard 4 on Amazon SageMaker AI with p5.48xlarge instances in the London AWS region
- 50+ Strands agents on Amazon ECS with RAG (pgvector) for document retrieval and agent orchestration
- Full UK data sovereignty for regulated industries like healthcare and legal, supporting ISO 42001 compliance
Why It Matters
Sets a blueprint for sovereign AI deployments in regulated sectors, proving self-hosted LLMs can scale securely on AWS.