Enterprise Azure-native GenAIOps platform featuring Retrieval-Augmented Generation (RAG), observability, Infrastructure as Code, CI/CD, containerized deployment, and cloud-native AI architecture.
This project demonstrates how modern enterprise AI systems are designed, deployed, monitored, and operated using Azure-native services.
The platform includes:
- Azure OpenAI integration
- Enterprise RAG using Azure AI Search
- FastAPI AI API service
- Docker containerization
- Azure Container Apps deployment
- OpenTelemetry tracing
- Application Insights observability
- Infrastructure as Code using Bicep
- GitHub Actions CI/CD
- Azure Key Vault security
- Managed Identity and RBAC
User / Client
↓
Azure Container Apps
↓
FastAPI Enterprise API
↓
Azure AI Search (RAG Retrieval)
↓
Azure OpenAI
↓
Grounded AI Response
↓
Application Insights + Log Analytics
GitHub Actions → CI/CD
Bicep + azd → Infrastructure Deployment
Azure Container Registry → Container Images
Key Vault → Secret Management
Managed Identity → Secure Authentication
| Service | Purpose |
|---|---|
| Azure OpenAI | LLM inference |
| Azure AI Search | Enterprise retrieval |
| Azure Container Apps | Cloud hosting |
| Azure Container Registry | Docker image storage |
| Application Insights | Monitoring and telemetry |
| Log Analytics | Centralized logging |
| Azure Key Vault | Secret management |
| Managed Identity | Passwordless authentication |
| GitHub Actions | CI/CD pipelines |
| Bicep | Infrastructure as Code |
| Azure Developer CLI (azd) | Environment orchestration |
- Retrieval-Augmented Generation (RAG)
- Cloud-native AI deployment
- Containerized AI APIs
- OpenTelemetry tracing
- AI smoke testing
- Infrastructure as Code
- RBAC-based security
- Managed Identity integration
- Azure observability stack
- GitHub Actions automation
- FastAPI API architecture
my-genaiops-agent/
│
├── agents/
│ ├── assistant_v1.py
│ ├── assistant_v4.py
│ └── assistant_v9.py
│
├── prompts/
│ ├── system_prompt_v1.txt
│ ├── system_prompt_v2.txt
│ └── system_prompt_v3.txt
│
├── evaluations/
│ ├── smoke_test.py
│ └── evaluation scripts
│
├── traces/
│ ├── trace_demo.py
│ └── azure_trace_demo.py
│
├── tools/
│ ├── memory.py
│ ├── logger.py
│ ├── keyvault_demo.py
│ └── azure_search_retriever.py
│
├── scripts/
│ ├── prompt_loader.py
│ ├── create_search_index.py
│ ├── upload_documents.py
│ └── query_search.py
│
├── infra/
│ └── main.bicep
│
├── .github/
│ └── workflows/
│ └── infra-validation.yml
│
├── azure.yaml
├── requirements.txt
└── .env
pip install -r requirements.txtuvicorn api.main:app --reloadhttp://127.0.0.1:8000/docs
docker build -t genaiops-api .docker run -p 8000:8000 --env-file .env genaiops-apiazd provisionazd deployGitHub Actions pipeline automates:
- dependency installation
- infrastructure validation
- smoke testing
- deployment workflows
The platform integrates:
- OpenTelemetry
- Azure Monitor
- Application Insights
- Log Analytics
This enables:
- tracing
- telemetry
- monitoring
- operational visibility
Security implementation includes:
- Azure Key Vault
- Managed Identity
- RBAC authorization
- GitHub Secrets
- secret scanning awareness
This project demonstrates practical experience with:
- GenAIOps
- Azure-native AI systems
- enterprise RAG
- cloud observability
- CI/CD pipelines
- Infrastructure as Code
- containerized deployments
- operational AI lifecycle management
Potential future enhancements:
- Vector search
- Semantic ranking
- Prompt evaluation gates
- FastAPI authentication
- Managed Identity for Key Vault access
- Multi-agent orchestration
- Azure dashboards
- Cost optimization
Azure-native GenAIOps learning platform featuring enterprise RAG, CI/CD, observability, IaC, Azure AI Search, Key Vault, and OpenTelemetry.
Hirdesh Pal