Advanced workshop designed to transition engineers from manual infrastructure management to Agentic DevOps workflows. It shifts the focus from traditional interactive AI to autonomous, context-aware systems, utilizing Claude Code as the primary engine for software lifecycle orchestration. Participants gain mastery in Context Engineering by codifying project standards and rules, effectively transforming LLMs into specialized team members capable of complex reasoning, multi-file manipulation, and independent task execution.

The curriculum emphasizes the architecture of multi-agent systems, integrating the Model Context Protocol (MCP) to bridge AI agents with external data sources, CI/CD pipelines, and cloud infrastructure. A critical component of the training is the implementation of “human-in-the-loop” safety guardrails, ensuring that while agents possess the agency to automate deployments and refactoring, they operate within strict, verifiable permission structures.

The Workshop culminates in the practical deployment of applications, demonstrating the ability to manage an AI-driven DevOps team in real-world production environments.

Objectives

• Implement Contextual Governance: Master the creation of CLAUDE.md documentation to enforce project-specific coding standards, architecture rules, and operational constraints for AI agents.
• Orchestrate Multi-Agent Systems: Manage parallel workflows by deploying subagents with distinct roles to perform complex, multi-stage DevOps tasks.
• Integrate Infrastructure via MCP: Utilize the Model Context Protocol (MCP) to securely connect AI agents to databases, remote infrastructure, and testing environments.
• Operationalize AI Safety: Configure robust “Hook” mechanisms, granular permissions, and approval workflows to prevent unintended infrastructure modifications.
• Automate CI/CD Pipelines: Execute end-to-end automated deployments, perform real-time troubleshooting, and manage code-base maintenance (PRs and bug fixes) using autonomous agentic tools.
• Adopt an Agentic Mindset: Shift from manual script execution to high-level command orchestration, allowing agents to reason through system-wide challenges.

Target Audience

• DevOps Engineers & Site Reliability Engineers (SREs): Need to transition from manual infrastructure management and script writing to orchestrating autonomous agentic workflows for automated deployments, continuous integration, and rapid system troubleshooting.
• Platform Engineers: Need to master Model Context Protocol (MCP) integration, configure secure hooks and permission boundaries, and embed AI automation safely into enterprise cloud infrastructure.
• Software Engineers & Full-Stack Developers: Need to accelerate their development velocity by adopting advanced context engineering (CLAUDE.md) and leveraging AI agents for multi-file refactoring, code maintenance, and automated pull request (PR) management.
• Technical Leads & Engineering Managers: Need to understand the architecture of multi-agent systems and safety guardrails to successfully implement and govern AI-driven development workflows across their teams.

Prerequisites & Experience

• Familiarity with DevOps: basic understanding of application deployment workflows and cloud environments.
• Software development background: basic coding literacy and familiarity with Git version control.
• Al fundamentals: basic understanding of large language models (LLMs) and prompt engineering.
• An active paid Claude plan with access to Claude; access must be verified before the workshop
• A personal computer for each participant

Agenda

How Agentic Al Works

A concise dive into the theory behind autonomous Al agents – how they reason, decide, and act differently from a standard chatbot.

Teaching Al Your Project

Master the CLAUDE.md file – the core of context engineering – to define your project’s rules, structure, and preferred coding style for the agent.

Claude Code Skills – Open Standard

Build advanced fluency with Claude Code’s built-in commands and slash commands and learn the open standards that let the agent integrate broadly across software projects.

First Deploy – Skills in Action

Your first real DevOps move: guide the agent through an initial deployment of code and infrastructure while monitoring and troubleshooting in real time.

Subagents – Your Al Team

Manage a multi-agent system: see how Claude Code spawns subagents, assigns them roles, and runs complex tasks in parallel –    just like a real dev team.

MCP Servers

The most comprehensive section in the course: connect your Al agent to external data sources, databases, testing tools, and local or remote cloud infrastructure using the Model Context Protocol (MCP).

Safety – Hooks, Permissions & Configuration

A critical section on safety and damage control: configure strict permissions and approval mechanisms (Hooks) for sensitive commands, so the agent never deletes or damages your infrastructure.

Real-World Workflows

Analyze and practice real-world case studies: automate bug fixes, manage pull requests (PRs), and streamline your Cl/CD pipeline.

Hands On – Deploy Your Application

A hands-on closing exercise: guide the agent team you built through a complete, end-to-end deployment of your application to the cloud.

The most comprehensive section in the course: connect your Al agent to external