OneAdvanced, a UK enterprise software provider, built a UK-sovereign AI platform on AWS, according to the AWS Machine Learning Blog. The company self-hosts the Llama 4 Maverick and Llama Guard 4 models on Amazon SageMaker AI, uses a retrieval-augmented generation pipeline built on pgvector, and has deployed more than 50 agents using the Strands Agents SDK running on Amazon ECS.
Why it matters
The deployment illustrates an approach to running AI workloads within a UK-sovereign environment while using open-weight models and managed cloud services. Combining self-hosted models with a RAG pipeline and multiple agents reflects a pattern for enterprises seeking control over data residency.
Who should care
Enterprises and technical teams evaluating sovereign AI deployments, agent frameworks, or self-hosting open-weight models on AWS may find the described architecture relevant.