All sectors
AI infrastructure audit and alternative cloud platform
The existing AI infrastructure was surveyed resource by resource, with the risks and the GPU line that accounted for most of the bill. A replacement platform hosted in France was built and tested end to end to give the client a costed option.
Context
An AI platform in production on a cloud shared with around twenty other applications.
The problem
High GPU costs and dependence on a foundation outside the client's control.
What we built
An API-based survey of 153 resources, a fresh-eyes report, costing, and a target platform on Scaleway built as infrastructure as code (Kubernetes cluster, L4 GPUs, managed databases), tested end to end with the embedding models.
Steps
- 01Survey and mapping
- 02Findings and costing
- 03Build of the target platform
- 04Migration plan and allocation of responsibilities
Hosting and models
Current: Azure West Europe. Target: Scaleway (France), vLLM and open models.
Services involved
AI & Cloud Strategy
We name a client only with their written agreement. The budgets and detailed results of our engagements remain confidential.