AI Engineering Team with Azure DevOps Integration 🤖➡️🚀

A revolutionary multi-agent AI engineering team that automatically processes Azure DevOps backlog items, generating complete full-stack applications with bilingual (English/Chinese) support, cost tracking, and detailed performance analytics.

What This Demo Shows

This project demonstrates how Agentic AI teams can integrate directly into Agile workflows:

• 🧑‍💻 Scrum roles as agents (Scrum Master, Product Owner, Engineer) collaborating on tasks
• 📊 Azure DevOps integration for backlog, task tracking, and status updates
• 💰 Token & cost usage tracking per agent role, with full visibility in Weights & Biases

🔗 Watch the demo: LinkedIn Post
🔗 Explore more projects: My GitHub Portfolio

⚠️ Stubbed Code Notice

This public repo demonstrates the architecture and CrewAI agent setup.
For security and commercial reasons, some integration logic has been stubbed out, including:

• ado_webhook_server.py → webhook listener, ADO event handling, cost tracking
• ado_client.py → Azure DevOps API integration (comments, status updates, cost posting)
• main.py → cost distribution and production-level orchestration

You can still run local demos (python main.py) to see CrewAI agents in action.

The full webhook → Azure DevOps → CrewAI → Weights & Biases integration is only available via OptimOps.ai.

📩 Contact: al@optimops.ai

🚀 Key Features

🤖 Multi-Agent AI Team

• Engineering Lead: Architect and requirements analysis
• Backend Engineer: Python module implementation
• Frontend Engineer: Gradio UI development
• Test Engineer: Comprehensive unit testing

🔄 Azure DevOps Native Integration

• Automatic triggering from ADO work items
• Real-time status updates and comments
• Cost transparency with detailed breakdowns
• Clarification protocol for ambiguous requirements

📊 Advanced Monitoring & Tracing

• Weave tracing for full execution visibility
• Weights & Biases integration for performance analytics
• Real-time token usage tracking
• Cost-per-task breakdowns

🌍 Enterprise Ready

• Bilingual support (English/Chinese)
• Cost-controlled development
• Azure DevOps native workflows
• Production-ready output

📱 Resulting App (Generated by the CrewAI Agents)

If everything goes well, your resulting app could look similar to this.
Every run, the LLMs get creative and may change the UI design,
but the core functionality for trading and transactions remains consistent.

🖼️ Screenshots

Main Page

Trading View

Portfolio Dashboard

Transaction History

▶️ Running the Sample Stock Trading App created by the Crew ***

The latest example generated by the CrewAI run is included in this repo under the /output folder.
The CrewAI-generated app uses a Gradio UI.

1. Navigate to the output folder:

   cd output
   
   

2. Start the Gradio app: python app.py

   If you prefer a specific port: python app.py --server_port 8000

3. Open your browser at: http://localhost:8000 or http://localhost:7860

🏗️ Architecture Overview

📸 Visual Demo
Azure DevOps Integration

Product Owners enter system requirements as a prompt in ADO

Real-Time Agent Updates

AI agents automatically process backlog items and post status updates

Weave Tracing & Analytics

Full execution tracings with Weave and performance analytics with W&B

Cost Transparency

Detailed cost breakdown per agent and model usage

🛠️ Quick Start Prerequisites

Python 3.10+

Azure DevOps organization

OpenAI API key

Deepseek API key (optional)

Installation bash

Clone the repository

git clone https://github.com/amateus1/crewai-engineering-team.git cd crewai-engineering-team

Checkout the stable branch

git checkout ado_improved_wandb

Install dependencies

pip install uv uv pip install -r requirements.txt

***sample requirements.txt *** pyyaml crewai sqlalchemy>=2.0.0 langchain openai

Environment Configuration bash

Copy and configure environment variables

cp .env.example .env

Add your API keys and ADO configuration

OPENAI_API_KEY=your_openai_key DEEPSEEK_API_KEY=your_deepseek_key ADO_ORG=your_azure_devops_org ADO_PROJECT=your_project_name ADO_PAT=your_personal_access_token

Azure DevOps Webhook Setup

Navigate to your ADO project settings

Create a new service hook with Webhook type

Set the URL to your server's webhook endpoint

Subscribe to Work item updated events

Filter for assigned to "Agentic Lead"

https://images/webhook-setup-screenshot.png Running the Server bash

Start the webhook server

python ado_webhook_server.py

Or run directly for testing

python -m engineering_team.main

🎯 How It Works

1. Backlog Item Creation

Create a work item in ADO with detailed requirements in the description: markdown

A simple account management system for a trading simulation platform. The system should allow users to create an account, deposit funds, and withdraw funds. The system should be bilingual in English and Chinese.

1. Automatic Processing

   AI team analyzes requirements

   Engineering Lead creates technical design

   Backend Engineer implements Python module

   Frontend Engineer builds Gradio UI

   Test Engineer writes comprehensive tests

2. Real-time Updates

   Status comments posted to ADO

   Cost breakdowns provided upon completion

   Weave traces generated for full visibility

3. Delivery

   Complete Python module delivered

   Functional UI application

   Comprehensive test suite

   Bilingual compliance report

📊 Performance Metrics Metric Value Description Avg. Development Time 2-5 minutes End-to-end task completion Cost per Task $0.10-$0.50 Based on complexity Token Usage 10K-50K Across all agents Success Rate 95%+ Production-ready output 🏢 Enterprise Features Cost Control python

Real-time cost tracking

PRICING = { "openai/gpt-4o": {"in": 2.50, "out": 10.00}, "deepseek/deepseek-chat": {"in": 0.14, "out": 0.28} }

Security & Compliance

No data persistence - all processing is ephemeral

Azure DevOps native - no external data storage

Enterprise-grade authentication with ADO PAT tokens

Scalability

Horizontal scaling with multiple worker instances

Queue processing for high-volume backlogs

Rate limiting to control API costs

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Fork the repository

Create a feature branch (git checkout -b feature/amazing-feature)

Commit your changes (git commit -m 'Add amazing feature')

Push to the branch (git push origin feature/amazing-feature)

Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details. 🆘 Support

Documentation: Check our docs

Issues: GitHub Issues

Discord: Join our community

Email: [Your email here]

🙏 Acknowledgments

crewAI for the amazing multi-agent framework

Azure DevOps for enterprise integration

Weights & Biases for tracing and analytics

OpenAI and Deepseek for LLM capabilities

📄 License

This project is released under the Business Source License 1.1 (BSL 1.1).
You are free to view the code, copy, modify, and use it for non-production purposes.
For any commercial or production use, please contact OptimOps.ai (amateus1) to obtain a separate license.

On 1 January 2030, this project will automatically transition to the GNU GPL v3 or later.

For full details, see the LICENSE file.

⭐ Star this repo if you find it useful!

Built with ❤️ by https://almateus.me and the power of AI collaboration text