LLMs and Agents in DevOps Workflows Training Course

Introduction to LLMs and Agent Frameworks

• Overview of large language models in infrastructure automation
• Core concepts in multi-agent workflows
• Applications of AutoGen, CrewAI, and LangChain in DevOps

Configuring LLM Agents for DevOps Tasks

• Installing AutoGen and setting up agent profiles
• Utilising the OpenAI API and other LLM providers
• Establishing workspaces and CI/CD-compatible environments

Automating Test and Code Quality Workflows

• Prompting LLMs to generate unit and integration tests
• Employing agents to enforce linting, commit rules, and code review standards
• Automating pull request summarisation and tagging

LLM Agents for Alert Handling and Change Detection

• Designing responder agents for pipeline failure alerts
• Analysing logs and traces using language models
• Proactively identifying high-risk changes or misconfigurations

Multi-Agent Coordination in DevOps

• Role-based agent orchestration (planner, executor, reviewer)
• Managing agent messaging loops and memory
• Incorporating human-in-the-loop designs for critical systems

Security, Governance, and Observability

• Managing data exposure and LLM safety within infrastructure
• Auditing agent actions and restricting operational scope
• Monitoring pipeline behaviour and model feedback

Real-World Use Cases and Custom Scenarios

• Developing agent workflows for incident response
• Integrating agents with GitHub Actions, Slack, or Jira
• Best practices for scaling LLM integration in DevOps

Summary and Next Steps