Hybrid-Cloud MLOps Automation Platform with AI Agents
🎥 Watch Demo Video - See the platform in action!
LLMOps Agent is an AI-powered MLOps platform that autonomously handles the complete ML lifecycle:
- 🤖 Intelligent model selection based on constraints (budget, time, performance)
- 📊 Automated dataset discovery from Hugging Face (100k+ datasets)
- 💰 Cost-optimized training Demo with SageMaker + LoRA
- 🔄 Hybrid cloud design (AWS now, on-prem later)
User Input:
"Train a Named Entity Recognition model on the ciER dataset. Budget: $10, Time: 1 hour, F1 score > 85%"
Agent Output (42 minutes later):
✅ Training complete! Model:
ner-ciER-distilbert-v1
📊 F1: 87.3%, Precision: 88%, Recall: 86%
💰 Cost: $4.20 (budget: $10.00)
⏱️ Time: 42 min (limit: 60 min)
graph TB
User[User] --> Orchestrator[Orchestrator Agent<br/>Bedrock AgentCore]
Orchestrator --> Data[Data Agent]
Orchestrator --> Model[Model Selection]
Orchestrator --> Train[Training Agent]
Data --> HF[Hugging Face]
Model --> Registry[Model Registry]
Train --> SageMaker[SageMaker]
- Python 3.11 or 3.12
- Poetry for dependency management
- Node.js 18+ (for frontend)
- AWS account with credits
git clone https://github.com/your-org/AgenticAIOps.git
cd AgenticAIOps# Copy environment template
cp .env.example .env
# Edit .env with your AWS credentials
# (See ENV_SETUP_COMPLETE.md for details)Backend:
# Install all dependencies (includes dev, docs, jupyter)
poetry install
# Or install only production dependencies
poetry install --only mainFrontend:
cd frontend
npm install
cd ..# Activate Poetry shell
poetry shell
# Verify environment configuration
poetry run llmops verify
# Or run directly
python -m llmops_agent.scripts.verify_envFollow the detailed guide: AWS Setup Instructions
Quick checklist:
- Apply $100 AWS credits
- Enable Bedrock (Claude 3.5 Sonnet)
- Create S3 buckets
- Create DynamoDB tables
- Configure IAM roles
Backend API:
# Start the FastAPI backend server on port 8003
poetry run uvicorn llmops_agent.api.main:app --reload --host 0.0.0.0 --port 8003
# Visit http://localhost:8003
# API Docs: http://localhost:8003/docsFrontend:
cd frontend
npm run dev
# Visit http://localhost:3000Documentation Server (optional):
poetry run mkdocs serve
# Visit http://localhost:8001# Run tests
poetry run pytest
# Format code
poetry run black src/ tests/
# Type checking
poetry run mypy src/
# View available commands
poetry run llmops --helpFor detailed Poetry usage, see POETRY_SETUP.md
Common commands:
# Add a dependency
poetry add package-name
# Add a dev dependency
poetry add --group dev package-name
# Update dependencies
poetry update
# Show installed packages
poetry show
# Export requirements.txt (for compatibility)
poetry export -f requirements.txt --output requirements.txtAgenticAIOps/
├── src/
│ └── llmops_agent/ # Main Python package
│ ├── api/ # FastAPI application
│ ├── agents/ # Agent implementations
│ ├── core/ # Core utilities
│ ├── models/ # Data models
│ ├── services/ # Business logic
│ ├── cli.py # CLI commands
│ └── config.py # Configuration management
├── tests/ # Test suite
│ ├── unit/ # Unit tests
│ └── integration/ # Integration tests
├── frontend/ # Next.js frontend
│ ├── app/ # Pages (chat, jobs, models, metrics)
│ ├── components/ # React components
│ ├── hooks/ # Custom hooks
│ └── scripts/ # Config sync scripts
├── scripts/ # Backend utility scripts
│ ├── setup_bedrock_agent.sh
│ ├── sync-ui-config.js
│ └── training/ # Training scripts
├── lambda/ # AWS Lambda functions
├── pyproject.toml # Poetry configuration
├── poetry.lock # Locked dependencies
├── .env # Environment variables (not committed)
└── .env.example # Environment template
Frontend: Next.js 14, TypeScript, Tailwind CSS, Radix UI Backend: FastAPI, Python 3.12, Poetry, Uvicorn Agents: Amazon Bedrock AgentCore, Claude 3.5 Sonnet, LangGraph ML: Hugging Face Transformers, PEFT (LoRA), SageMaker, PyTorch MLOps: MLflow, S3, DynamoDB Monitoring: CloudWatch, Structured Logging Package Management: Poetry, npm
Event: AWS AI Agent Global Hackathon
Deadline: October 22, 2025
Submission: Working POC + 3-min demo video
- Frontend UI (Next.js)
- Architecture design
- Documentation setup
- Backend implementation (FastAPI)
- Bedrock AgentCore integration
- SageMaker training pipeline
- End-to-end training automation
- Demo video (Watch here)
- Frontend Local Setup & Testing Support
- Backend Local Setup & Testing Support
- OnPrem MLOps Support
- Multiagentic MLOps Improvements
- Multimodal AIOps Improvements
Currently in hackathon mode. Contributions welcome reachout to the authors.
MIT (to be finalized)
Developer: Sri Chakra, Manu Chandran
Hackathon: AWS AI Agent Global Hackathon 2025