Operate

AI Workload Center

Reference implementation patterns for the workloads that run on GPU platforms — architecture, stack, sizing, and a sample manifest for each.

Training
Training

Distributed foundation-model training

Synchronous data- and model-parallel training across many GPUs, bounded by the collective-communication fabric.

Open pattern Training

Parameter-efficient fine-tuning (LoRA/QLoRA)

Adapt a base model with low-rank adapters — a fraction of the compute and memory of full fine-tuning.

Open pattern
Inference
Inference

Real-time LLM inference

Low-latency, autoscaling serving with continuous batching behind an API gateway.

Open pattern Inference

Batch / offline inference

Throughput-first offline scoring or embedding generation, ideal for spot capacity.

Open pattern
Retrieval / RAG
Retrieval / RAG

Retrieval-augmented generation (RAG)

Ground model responses in enterprise knowledge via embedding, retrieval, reranking, and generation.

Open pattern
Multimodal & vision
Multimodal & vision

Multimodal & computer vision

Image/video understanding and multimodal models — vision encoders plus generation.

Open pattern
Agentic
Agentic

Agentic AI & multi-agent systems

LLM agents that plan and call tools — the hard parts are tool safety, isolation, and cost control.

Open pattern
HPC & scientific
HPC & scientific

HPC & scientific computing (Slurm + MPI)

Tightly-coupled simulation — CFD, molecular dynamics, weather, genomics — run as MPI jobs, often under Slurm alongside or bridged to Kubernetes.

Open pattern