AI Demos Collection

A comprehensive collection of enterprise AI applications and demonstrations for SUSE's cloud-native platform

This repository serves as a curated catalog of AI-related applications and demonstrations, showcasing deployment, security, observability, and management using SUSE's enterprise-grade cloud-native stack. Each application is packaged as a production-ready Helm chart, deployable via Rancher Apps & Marketplace, Helm CLI, or GitOps with Fleet.

---

📋 Table of Contents

• 🎯 Repository Overview
• 🚀 AI Compare Application
• 📖 Available Demonstrations
• 🛠️ Quick Start
• 📁 Repository Structure
• 🔧 Advanced Configuration
• 🤝 Contributing

---

🎯 Repository Overview

This repository provides a curated collection of enterprise-ready AI applications and comprehensive demonstration materials organized into three main categories:

1. 🤖 AI Applications Catalog

Production-ready Helm charts for AI workloads, each available in multiple variants:

• AI Compare - AI response comparison tool with security and observability demonstrations
  • ai-compare-suse: Enterprise SUSE BCI-based edition
  • ai-compare: Upstream community edition
  • ai-compare-opentelemetry: Advanced GenAI observability edition with token/cost tracking
• Ollama - Local LLM inference server with GPU acceleration
  • ollama-suse: SUSE enterprise edition
  • ollama-upstream: Upstream community edition
  • ollama-suse-direct: Direct NVIDIA GPU access variant

Key Application Features:

• Security Demonstrations: Built-in NeuVector DLP testing with dual data type transmission
• Enterprise Integration: OpenTelemetry observability, GPU acceleration, persistent storage
• Multi-Deployment Options: Rancher UI, Helm CLI, GitOps with Fleet
• Flexible Architecture: Direct model access or pipeline-enhanced processing

2. 📋 Platform Demonstrations

Guided demonstrations covering the complete AI application lifecycle:

• Infrastructure: GPU provisioning and management with Rancher
• Deployment: Multiple deployment methodologies for AI workloads
• Observability: AI-specific monitoring with SUSE Observability
• Security: Container security and runtime protection with NeuVector
• Zero-Trust: Network security and policy enforcement

3. 🏗️ Enterprise Infrastructure

Complete deployment automation and configurations:

• Helm Charts: Production-ready charts for all applications with SUSE and upstream variants
• GitOps: Fleet-based continuous deployment configurations
• Observability: Pre-configured monitoring and alerting
• Security: NeuVector policy automation and DLP configurations

📦 Available Charts

Chart Name	Version	Description	Variants
ai-compare-suse	0.1.x	Enterprise AI comparison app (SUSE BCI)	Production
ai-compare	0.1.x	AI comparison app (Upstream)	Community
ai-compare-opentelemetry	0.1.x	Enhanced GenAI observability edition	Advanced
ollama-suse	0.1.x	LLM inference server (SUSE)	Enterprise
ollama-upstream	0.1.x	LLM inference server (Upstream)	Community

Adding New Applications: This repository is designed to grow as a catalog of AI applications. To add new applications, package them as Helm charts and submit via pull request following the existing chart structure.

---

🚀 AI Compare Application

Core Functionality

The flagship AI Compare application provides real-time comparison between:

• 🤖 Direct Ollama: Local LLM inference (TinyLlama, Llama2, custom models)
• 🌐 Pipeline-Enhanced: Processed responses through Open WebUI pipelines with educational levels:
  • 👶 Kid-friendly explanations
  • 🎓 Student-level responses
  • ⚗️ Scientific detailed analysis

Built-in Security Demonstrations

🔒 Data Leak Prevention (DLP) Demo

• Dual Data Transmission: Single button sends both credit card and SSN data
• Credit Card: 3412-1234-1234-2222
• Social Security Number: 123-45-6789
• NeuVector Integration: Triggers real-time DLP monitoring and alerting
• Clean Interface: Simple popup showing "⚠️ Attempting to send sensitive data"

🌐 Availability Demo

• External Connectivity: Tests connection to https://suse.com
• Network Policy Validation: Demonstrates network segmentation capabilities
• Security Monitoring: Validates outbound connection policies

Enterprise Features

• 📊 Real-time Observability: OpenTelemetry integration with SUSE Observability
• 🖥️ GPU Acceleration: NVIDIA GPU support with runtime configuration
• 💾 Persistent Storage: Model caching and configuration persistence
• 🔄 Automation: Background testing and response comparison
• 👥 Provider Monitoring: Live status of major AI providers (OpenAI, Anthropic, Google, etc.)

---

📖 Available Demonstrations

Platform Demonstrations (Guided Walkthroughs)

Demo	Focus Area	Duration	Key Takeaways
Demo 1: Accelerating AI with Rancher and GPUs	Infrastructure	15 min	GPU provisioning, cluster management, hardware optimization
Demo 2: Deploying the SUSE AI Stack	Deployment	20 min	Rancher UI, Helm CLI, GitOps deployment methods
Demo 3: Monitoring AI with SUSE Observability	Observability	15 min	GPU metrics, cost tracking, performance optimization
Demo 4: Building Trustworthy AI	Security	20 min	Container scanning, vulnerability management, policy automation
Demo 5: Zero-Trust Security for AI	Network Security	15 min	Runtime protection, network policies, threat detection

Interactive Application Demos (Built-in)

Demo	Trigger	Data Transmitted	NeuVector Detection
🔒 Data Leak Demo	Single Button	Credit Card + SSN	Multi-pattern DLP alerts
🌐 Availability Demo	Single Button	HTTPS Request	Network policy validation

---

🛠️ Quick Start

Prerequisites

• Kubernetes cluster (RKE2 recommended)
• Helm 3.x
• kubectl configured
• Optional: GPU nodes with NVIDIA drivers

Adding the Repository

Option 1: Rancher UI (ClusterRepo)

Add this repository to Rancher's Apps & Marketplace by creating a ClusterRepo resource:

1. Navigate to Cluster → More Resources → Catalog (catalog.cattle.io) → ClusterRepos
2. Click Create from YAML
3. Apply the following configuration:

apiVersion: catalog.cattle.io/v1
kind: ClusterRepo
metadata:
  name: ai-demos
spec:
  url: https://wiredquill.github.io/ai-demos

4. Click Create - Charts will appear in Apps & Marketplace within 1-2 minutes

Option 2: Helm Repository

# Add the Helm repository
helm repo add ai-demos https://wiredquill.github.io/ai-demos

# Update repository index
helm repo update

# Search available charts
helm search repo ai-demos

# Install a chart
helm install my-release ai-demos/ai-compare-suse

Deployment Options

Option 1: Rancher Apps & Marketplace (Recommended)

1. Open Rancher cluster management interface
2. Navigate to Apps & Marketplace
3. Search for "AI Compare"
4. Select variant:
   • ai-compare-suse: Enterprise SUSE edition
   • ai-compare: Upstream community edition
5. Configure and deploy

Option 2: Helm CLI

# SUSE Enterprise Edition
helm install ai-demo charts/ai-compare-suse \
  --set ollama.gpu.enabled=true \
  --set aiCompare.observability.enabled=true

# Upstream Community Edition
helm install ai-demo charts/ai-compare

Option 3: GitOps with Fleet

# Deploy Fleet configuration
kubectl apply -f fleet/fleet.yaml

# Label target clusters
kubectl label cluster my-cluster needs-llm-suse=true

Access the Application

# Port forward to access locally
kubectl port-forward svc/ai-demo-app-service 7860:7860

# Open browser to http://localhost:7860

---

📁 Repository Structure

ai-demos/
├── 📱 app/                           # AI Compare application source
│   ├── python-ollama-open-webui.py     # Main Gradio application
│   ├── Dockerfile.suse                 # SUSE BCI-based container
│   ├── Dockerfile.upstream             # Debian-based container
│   ├── requirements.txt                # Python dependencies
│   └── tests/                          # Application test suite
├── 📦 charts/                        # Production Helm charts
│   ├── ai-compare/                     # Upstream community chart
│   └── ai-compare-suse/               # SUSE enterprise chart
├── 📋 Demo Guides/                   # Step-by-step demonstrations
│   ├── demo-1.md                       # GPU infrastructure with Rancher
│   ├── demo-2.md                       # Multi-method deployment
│   ├── demo-3.md                       # SUSE Observability monitoring
│   ├── demo-4.md                       # Container security with NeuVector
│   └── demo-5.md                       # Zero-trust network security
├── 🚀 fleet/                         # GitOps deployment automation
│   ├── fleet.yaml                      # Fleet configuration
│   └── gpu-operator/                   # GPU operator automation
├── 🔧 install/                       # Infrastructure setup guides
│   ├── README.md                       # Installation overview
│   ├── Install-GPU-Operator.md        # GPU infrastructure setup
│   ├── Enable-SUSE-AI-Observability.md # Monitoring configuration
│   └── Install-NVIDIA-drivers.md      # Driver installation guide
├── 🔄 pipelines/                     # AI pipeline configurations
│   ├── response_level_pipeline.py     # Educational response processing
│   └── pipeline_config.yaml           # Pipeline configuration
├── 📊 docs/                          # Technical documentation
│   ├── OPENTELEMETRY-INTEGRATION.md   # Observability setup
│   ├── AUTOMATED-DEPLOYMENT.md        # CI/CD configuration
│   └── AI-MODEL-CACHING.md           # Model caching strategies
├── 🖼️ assets/                        # Screenshots and documentation images
└── 🔨 scripts/                       # Automation and utility scripts

---

🔧 Advanced Configuration

GPU Acceleration

ollama:
  gpu:
    enabled: true
  hardware:
    type: nvidia

Enterprise Observability

aiCompare:
  observability:
    enabled: true
    otlpEndpoint: "http://opentelemetry-collector.suse-observability.svc.cluster.local:4318"
    collectGpuStats: true

Security Integration

neuvector:
  enabled: true
  dlpPolicies: true
  securityDemos: true

Development Mode

aiCompare:
  devMode:
    enabled: true
    persistence:
      enabled: true
    gitRepo: "https://github.com/your-org/ai-demos.git"

---

🤝 Contributing

This repository welcomes contributions of new AI applications, improved demonstrations, and enhanced documentation.

Adding New AI Applications to the Catalog

To add a new AI application to the repository:

1. Create Helm Chart Structure

   # Create chart directory following naming convention
   mkdir -p charts/your-app-name-suse
   mkdir -p charts/your-app-name  # for upstream variant

2. Package Your Application

   • Follow existing chart patterns (see charts/ai-compare-suse as reference)
   • Include both SUSE BCI and upstream variants when possible
   • Add comprehensive values.yaml with documentation
   • Include README.md explaining application purpose and configuration

3. Test Your Chart

   # Lint the chart
   helm lint charts/your-app-name-suse
   
   # Test deployment
   helm install test-release charts/your-app-name-suse
   helm test test-release

4. Submit Pull Request

   • Charts are automatically packaged and published to gh-pages branch via CI/CD
   • Include demo documentation if applicable
   • Update main README.md to list your application in the catalog table

General Contributions

1. Fork the repository
2. Create a feature branch: git checkout -b feature/new-app or git checkout -b feature/improved-demo
3. Make your changes: Add new applications, demos, or improvements
4. Test thoroughly: Ensure all changes work in both SUSE and upstream environments
5. Submit a pull request: Include clear description of changes and testing performed

Development Environment

# Setup local development
git clone https://github.com/wiredquill/ai-demos.git
cd ai-demos

# For application development
cd app
pip install -r requirements.txt
python python-ollama-open-webui.py

# For chart development
helm lint charts/your-chart-name
helm install test charts/your-chart-name --dry-run --debug

# Run tests
pytest tests/
helm test my-release

Chart Packaging and Publishing

Charts are automatically packaged and published via GitHub Actions:

• Commits to main trigger automatic chart packaging
• Charts are published to gh-pages branch
• Rancher ClusterRepo and Helm repository automatically pick up updates
• Manual packaging: See .github/workflows/ for CI/CD pipeline details

---

📜 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

🆘 Support

• 📖 Documentation: Check the docs/ directory for detailed technical guides
• 🛠️ Installation Help: Review install/README.md for setup assistance
• 🐛 Issues: Open an issue in this repository for bug reports or feature requests
• 💬 Discussions: Use GitHub Discussions for questions and community support

---

Powered by SUSE's Enterprise Cloud-Native AI Platform

Complete demonstrations of enterprise AI workloads from infrastructure provisioning to application security and observability.