| name | oracle-ai-blueprints |
| description | OCI AI Blueprints and AI Accelerator Packs — no-code platform for deploying GenAI workloads on OKE. Check before designing custom GenAI infrastructure. |
| version | 1.0.0 |
| platform | ["claude-code","cline","cursor","roocode"] |
| activation | {"cline":"@skills/oracle-ai-blueprints/SKILL.md","cursor":"@skills/oracle-ai-blueprints/SKILL.md"} |
OCI AI Blueprints & Accelerator Packs
You are an expert in OCI AI Blueprints — the no-code platform for deploying GenAI workloads on OKE.
CRITICAL: Check AI Blueprints BEFORE designing custom GenAI infrastructure.
Repository: https://github.com/oracle-quickstart/oci-ai-blueprints
Available Blueprints
| Blueprint | Use Case | Key Features |
|---|
| LLM Inference (vLLM) | Deploy Llama/Mistral | NVIDIA GPU shapes, auto-scaling |
| Llama Stack | Complete GenAI runtime | vLLM + ChromaDB + Postgres + Jaeger |
| LoRA Fine-Tuning | Custom model training | HuggingFace models, hyperparameter tuning |
| Multi-node Inference | Llama-405B sized | RDMA + H100 nodes + LeaderWorkerSet |
| Autoscaling Inference | Production serving | KEDA + latency-based scaling |
| MIG Inference | Cost optimization | Multi-instance GPU partitioning |
| CPU Inference | Budget/testing | Ollama + Mistral/Gemma |
AI Accelerator Packs (Enterprise ORM Stacks)
| Pack | Industry | NVIDIA Tech | GPU |
|---|
| Route Optimizer | Logistics | cuOPT | A100 |
| Video Intelligence | Media/Enterprise | Cosmos + Parakeet NIMs | A100 |
| AI-Q Self-Hosted | Enterprise (sovereign) | NVIDIA NIMs | 16x A100 |
| AI-Q with OCI GenAI | Enterprise (managed) | Llama Stack + GenAI | CPU + GenAI |
| Content Moderation | Media platforms | Multimodal analysis | A100 |
All packs include: Prometheus, Grafana, PostgreSQL, KEDA, MLFlow
Tech Stack
┌─────────────────────────────────────────────┐
│ AI ACCELERATOR PACK │
│ ┌─────────────────────────────────────┐ │
│ │ Application Layer │ │
│ │ Route Optimizer │ VSS │ AI-Q │ │
│ └─────────────────────────────────────┘ │
├─────────────────────────────────────────────┤
│ NVIDIA SOFTWARE │
│ cuOPT │ Cosmos │ NVIDIA NIMs │
├─────────────────────────────────────────────┤
│ INFERENCE LAYER │
│ vLLM (OpenAI-compatible API) │
├─────────────────────────────────────────────┤
│ OBSERVABILITY │
│ Prometheus │ Grafana │ MLFlow │ KEDA │
├─────────────────────────────────────────────┤
│ OCI INFRASTRUCTURE │
│ OKE │ GPU Compute │ Object Storage │
└─────────────────────────────────────────────┘
Customer Decision Tree
Customer needs GenAI on OCI?
├── Standard inference (Llama, Mistral)?
│ └── USE: AI Blueprints - vLLM or Llama Stack
├── Fine-tuning required?
│ └── USE: AI Blueprints - LoRA Fine-Tuning
├── Route optimization?
│ └── USE: AI Accelerator Pack - cuOPT
├── Video analysis?
│ └── USE: AI Accelerator Pack - VSS
├── Enterprise reasoning agent?
│ └── USE: AI Accelerator Pack - AI-Q
├── Multimodal / OCI GenAI Service?
│ └── USE: Cohere Command A Vision + GenAI endpoints
└── Custom/unique requirements?
└── Build custom with oracle-devrel patterns
Industry Fit Matrix
| Industry | Route Optimizer | Video Intelligence | AI-Q | GenAI Search |
|---|
| Logistics | +++ | + | ++ | +++ |
| Healthcare | ++ | +++ | +++ | +++ |
| Financial Services | + | + | +++ | +++ |
| Manufacturing | +++ | +++ | ++ | ++ |
| Media | + | +++ | ++ | ++ |
References
Cline Activation
To use this skill in Cline, reference it at the start of your message:
@skills/oracle-ai-blueprints/SKILL.md
A customer needs to deploy Llama 4 Maverick on OCI for a production RAG system. What blueprint should they use?
Or in a .clinerules workflow:
## GenAI Infrastructure
Before recommending custom GenAI infrastructure on OCI, load @skills/oracle-ai-blueprints/SKILL.md and check whether an existing blueprint or accelerator pack covers the requirement.
Triggers: OCI AI Blueprints, AI Accelerator Packs, vLLM OCI, LLM on OKE, Llama Stack OCI, route optimizer OCI, video intelligence OCI, AI-Q
