Comprehensive demo repository for AI Agent technologies with multi-framework examples (CrewAI, kagent, ADK), MCP integration, fine-tuning, and production-ready AI Gateway deployments.
Overview#
agentic-demo-repo is an AI Agent technology demo repository maintained by AdminTurnedDevOps, designed to lower the barrier for learning and production deployment of AI Agent technologies.
Core Capabilities#
Agent Framework Support#
- CrewAI: Multi-agent collaboration framework
- kagent: Kubernetes-native Agent framework (OSS/Enterprise)
- ADK (Agent Development Kit): Universal agent development toolkit
- LangChain: With Anthropic integration
- NVIDIA NeMo: GPU-accelerated inference
- AWS Bedrock / Bedrock AgentCore: AWS managed LLM
- Serverless ADK: Serverless deployment
AI Gateway Features#
- agentgateway Enterprise/OSS editions
- Kubernetes-native deployment (CRDs + Helm)
- Multi-LLM provider proxy (Anthropic, OpenAI, etc.)
- Gateway API v1.4.0 integration
MCP (Model Context Protocol)#
- MCP server and tool integration examples
Infrastructure Components#
- EKS Terraform configurations
- Helm Charts package management
- Keycloak authentication integration
- RBAC security policies
- Chaos Mesh testing
Model Fine-tuning#
- DeepSeek fine-tuning examples
- LLM testing and evaluation framework
Repository Structure#
agentic-demo-repo/
├── adk/ # ADK examples
├── agentgateway-*/ # Enterprise/OSS/K8s deployments
├── frameworks/ # Multi-framework examples
├── mcp/ # MCP integration
├── fine-tuning/ # Fine-tuning examples
├── eks-terraform/ # EKS infrastructure
├── security-agw-kagent/ # Security configurations
└── production-demos/ # Production demos
Quick Deployment#
Prerequisites#
- Kubernetes cluster (EKS recommended)
- Helm 3
- kubectl
- LLM provider API Key
Deployment Steps#
- Install kagent CRDs
- Configure API Key and install kagent
- Install Gateway API v1.4.0
- Install agentgateway
Key CRDs#
| CRD | Source | Purpose |
|---|---|---|
| Agent | kagent.dev/v1alpha2 | Define Agent instances |
| AgentgatewayBackend | agentgateway.dev/v1alpha1 | Define AI backend config |
Use Cases#
- Agent framework learning and comparison
- Production deployment reference
- AI Gateway POC
- Enterprise security architecture
- Multi-cloud/hybrid deployment
Language Composition#
Python (45.5%), Shell (38.5%), HCL (14.4%), Dockerfile (1.6%)
Pending Confirmation#
- Specific open source license type
- Official documentation site
- Community support channels