https://github.com/nasim-raj-laskar/polyaiagent-pipeline
Multi-agent LLM platform featuring Jenkins pipelines, Docker containerization, and AWS Fargate orchestration.
https://github.com/nasim-raj-laskar/polyaiagent-pipeline
aiops aws ci-cd llmops
Last synced: 3 months ago
JSON representation
Multi-agent LLM platform featuring Jenkins pipelines, Docker containerization, and AWS Fargate orchestration.
- Host: GitHub
- URL: https://github.com/nasim-raj-laskar/polyaiagent-pipeline
- Owner: nasim-raj-laskar
- Created: 2025-09-19T03:32:53.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2025-09-23T14:41:46.000Z (10 months ago)
- Last Synced: 2025-09-23T14:48:18.873Z (10 months ago)
- Topics: aiops, aws, ci-cd, llmops
- Language: Python
- Homepage:
- Size: 403 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# PolyAiAgent-pipeline
Production-grade multi-agent LLM orchestration platform with enterprise CI/CD pipeline and cloud-native deployment.
## Infrastructure Architecture
- **Compute**: AWS ECS Fargate serverless containers
- **Registry**: AWS ECR private container registry
- **CI/CD**: Jenkins with Docker-in-Docker (DinD) capability
- **Code Quality**: SonarQube static analysis integration
- **Runtime**: FastAPI REST API + LangGraph ReAct agents
## CI/CD Pipeline
### Jenkins Pipeline Stages
1. **SCM Integration**: GitHub webhook-triggered builds
2. **Static Analysis**: SonarQube code quality gates
3. **Container Build**: Multi-stage Docker builds with layer caching
4. **Registry Push**: AWS ECR image versioning and tagging
5. **Deployment**: ECS Fargate rolling updates with zero downtime
### Infrastructure Components
```
├── Jenkinsfile # Declarative pipeline configuration
├── custom_jenkins/Dockerfile # Jenkins DinD container setup
├── Dockerfile # Multi-stage application container
└── app/ # Application source code
```
### AWS Services Integration
- **ECS Cluster**: `multi-aiagent-cluster`
- **ECS Service**: `multi-ai-agent-service-1`
- **ECR Repository**: `multiaiagent`
- **Region**: `us-east-1`
## Application Stack
- **LLM Models**: `llama-3.3-70b-versatile`, `openai/gpt-oss-120b`
- **Agent Framework**: LangGraph ReAct pattern
- **API Gateway**: FastAPI with Pydantic validation
- **Search Integration**: Tavily API
## Deployment
### Production (AWS ECS Fargate)
```bash
# Triggered via Jenkins pipeline
aws ecs update-service --cluster multi-aiagent-cluster \
--service multi-ai-agent-service-1 --force-new-deployment
```
### Local Development
```bash
docker build -t multiaiagent .
docker run -p 8000:8000 -p 8501:8501 multiaiagent
```
## API Specification
```http
POST /chat
Content-Type: application/json
{
"model_name": "llama-3.3-70b-versatile",
"system_prompt": "You are a helpful assistant",
"messages": ["Hello"],
"allow_search": false
}
```
## Environment Configuration
```bash
GROQ_API_KEY=
TAVILY_API_KEY=
AWS_REGION=us-east-1
ECR_REPO=multiaiagent
```
## Complete Architecture
