private-ai-vcf · 4 articles
Private AI on VCF
A practical path for running private AI on VCF, from architecture to GPU pooling, vector databases and production RAG.
Start the series
Every article
Private AI on VCF: the architecture that fits in your DC
VCF Private AI Foundation is the stack Broadcom proposes for hosting LLMs and inference in your own DC. We dissect it, layer by layer, seams visible.
GPU pooling on VCF: NVIDIA vGPU + MIG in practice
Sharing GPUs across tenants is the only way private AI math works. We walk through vGPU profiles, MIG slicing, and the pitfalls of mixed-workload scheduling.
Vector databases on VKS: pgvector, Milvus, Weaviate
RAG needs a vector store. We compare pgvector, Milvus and Weaviate on VKS — index quality, ops surface, and which one actually fits your team.
RAG in production: from POC to scale on VCF
Most RAG POCs die between demo and production. We walk through the gap — chunking, freshness, evals, observability — on VCF and VKS.



