AI Architect Command Center

Transform Claude Code into your expert AI Architecture assistant specializing in Multi-Cloud AI solutions with Oracle Cloud Infrastructure focus.

Quick Start

# Clone and enter the repository
cd claude-ai-architect

# Start Claude Code with this as the working directory
claude

# Use slash commands to get started
/design-solution "Customer support AI with RAG"
/deploy-genai
/draw-architecture

What This Repository Does

This repository configures Claude Code as an AI Architect with expertise in:

• OCI GenAI & Dedicated AI Clusters - Design, deploy, and optimize GenAI solutions
• Multi-Cloud AI Architecture - AWS, Azure, GCP, OCI workload placement
• RAG & Agent Systems - Knowledge bases, retrieval, multi-agent orchestration
• Architecture Diagramming - Professional D2, Draw.io, Mermaid diagrams
• Enterprise AI Patterns - Security, governance, FinOps

Repository Structure

claude-ai-architect/
├── CLAUDE.md                    # Claude Code instructions
├── README.md                    # This file
│
├── knowledge-base/              # Domain expertise
│   ├── oci-genai/              # OCI GenAI documentation
│   ├── multi-cloud/            # Multi-cloud patterns
│   └── ai-infrastructure/      # GPU and infrastructure
│
├── skills/                      # 14 specialized skills
│   ├── architecture-diagramming/
│   ├── genai-dac-specialist/
│   ├── rag-expert/
│   ├── enterprise-ai-patterns/
│   └── ... (10 more)
│
├── templates/
│   ├── d2/                     # D2 diagram templates
│   ├── terraform/              # IaC modules
│   │   ├── oci-genai-dac/
│   │   ├── oci-rag-system/
│   │   ├── ai-gateway/
│   │   └── security-baseline/
│   ├── cost-calculator.md
│   └── discovery-questions.md
│
├── prompts/                    # Slash command definitions
│   ├── design-solution.md
│   ├── deploy-genai.md
│   ├── build-rag.md
│   ├── multi-cloud-setup.md
│   ├── optimize-costs.md
│   ├── security-review.md
│   └── draw-architecture.md
│
├── examples/                   # Real-world solutions
│   ├── customer-support-ai.md
│   └── multi-cloud-analytics.md
│
├── sdk-examples/               # Python code examples
│   └── oci_genai_examples.py
│
├── cheatsheets/               # Quick reference
│   ├── OCI-GENAI-CHEATSHEET.md
│   ├── D2-DIAGRAM-CHEATSHEET.md
│   └── MODEL-COMPARISON-MATRIX.md
│
├── mcp-servers/               # Custom MCP servers
│   ├── oci-infrastructure/    # OCI operations
│   └── terraform-ops/         # Terraform operations
│
├── agent-teams/               # Agent organization
├── workflows/                 # Architecture workflows
└── config/                    # MCP server configs

Slash Commands

Command	Description
/design-solution	End-to-end AI solution design
/deploy-genai	Deploy OCI GenAI Dedicated AI Cluster
/build-rag	Build RAG system with knowledge bases
/multi-cloud-setup	Configure multi-cloud AI architecture
/optimize-costs	Cost optimization review
/security-review	Security assessment
/draw-architecture	Generate architecture diagrams

Features

1. Knowledge Base

Comprehensive documentation on:

• OCI Dedicated AI Clusters (DACs)
• OCI GenAI Agents and RAG
• Multi-cloud AI patterns
• GPU infrastructure and training

2. Terraform Modules

Ready-to-deploy infrastructure:

cd templates/terraform/oci-genai-dac
terraform init && terraform apply

3. D2 Diagram Templates

Professional architecture diagrams:

d2 --layout=tala templates/d2/oci-genai-rag.d2 output.svg

4. Python SDK Examples

Working code for OCI GenAI:

# See sdk-examples/oci_genai_examples.py
from oci.generative_ai_inference import GenerativeAiInferenceClient

5. MCP Servers

Custom servers for Claude integration:

• oci-infrastructure - OCI API operations
• terraform-ops - Terraform plan/apply/state

Installation

Prerequisites

• Claude Code CLI installed
• OCI CLI configured (for OCI operations)
• Terraform installed (for IaC)
• D2 installed (for diagrams)

Install D2 (for diagrams)

# macOS
brew install d2

# Linux
curl -fsSL https://d2lang.com/install.sh | sh

# Windows (PowerShell)
iwr -useb https://d2lang.com/install.ps1 | iex

Install MCP Servers

# OCI Infrastructure Server
cd mcp-servers/oci-infrastructure
pip install -r requirements.txt

# Terraform Operations Server
cd mcp-servers/terraform-ops
pip install -r requirements.txt

Configure MCP Servers

Add to ~/.claude/settings.json:

{
  "mcpServers": {
    "oci": {
      "command": "python",
      "args": ["/path/to/mcp-servers/oci-infrastructure/server.py"]
    },
    "terraform": {
      "command": "python",
      "args": ["/path/to/mcp-servers/terraform-ops/server.py"]
    }
  }
}

Usage Examples

Design a Solution

User: Design an AI solution for customer support

Claude: I'll use the /design-solution workflow...
[Generates architecture, D2 diagram, Terraform, cost estimate]

Deploy GenAI Cluster

User: Deploy a production GenAI cluster

Claude: I'll use the /deploy-genai workflow...
[Recommends sizing, generates Terraform, provides validation steps]

Generate Diagram

User: Create an architecture diagram for our RAG system

Claude: I'll generate a D2 diagram...
[Creates professional diagram with OCI components]

Model Comparison

Model	Best For	Cost
Command R+	Complex reasoning, RAG	$$$
Command R	General purpose	$$
Command Light	High volume, simple	$
Llama 3.1 70B	Open source	$$$
GPT-4 Turbo	Best capability	$$$$
Claude 3.5 Sonnet	Best balance	$$$

See cheatsheets/MODEL-COMPARISON-MATRIX.md for full comparison.

Cost Reference

DAC Size	Units	Monthly Est.
Dev/Test	2-5	$3-7.5K
Production	5-15	$7.5-22.5K
High Volume	15-30	$22.5-45K
Enterprise	30-50	$45-75K

See templates/cost-calculator.md for detailed scenarios.

Contributing

1. Add knowledge to knowledge-base/
2. Create skills in skills/
3. Add templates to templates/
4. Document in cheatsheets/

Resources

OCI Documentation

• OCI GenAI
• Dedicated AI Clusters
• GenAI Agents
• OCI Icons

Diagramming

• D2 Language
• D2 Playground

Agent Frameworks

• Claude Agent SDK
• Oracle ADK
• MCP Protocol

License

MIT License - Use freely for your AI architecture projects.

---

AI Architect Command Center - Design with excellence, diagram with clarity, deploy with confidence.