About the course
While architectural theory is essential, the true challenge lies in the implementation. This 4-day "Deep Dive" moves from the drawing board to the IDE, focusing on the practical engineering of agentic systems using flagship tools like Claude Code, GitHub Copilot, and Codex.
We treat AI agents as first-class citizens in the development lifecycle. Participants will learn to build custom "Skills" via open standards, connect agents to internal enterprise tools via MCP servers, and manage a "team" of specialized subagents. This course is highly tactical, covering everything from supply chain security for AI-generated code to zero-trust authentication for non-human identities.
Customisable Delivery: This course is Part 2 of our Agentic Development Programme. It can be delivered as a standalone workshop for teams with existing architectural knowledge or as a follow-on to our Enterprise Agentic Architecture Foundations course. This course is the first part of our 2-week Agentic Development Programme.
-
- Build Extensible Agents: Develop custom Skills and MCP servers for enterprise tool integration.
- Orchestrate Agent Teams: Coordinate specialized subagents (Architect, Reviewer, Security).
- Harden the Pipeline: Integrate agents into CI/CD with auto-linting and test-failure remediation.
- Secure the Supply Chain: Protect against package hallucinations and dependency attacks.
- Manage Agent Identity: Implement scoped identities and audit trails for agent commits.
-
This workshop is aimed at:
Full-Stack Developers & DevOps Engineers looking to integrate AI agents into their active development workflows.
AI Engineers who need to build custom extensions (Skills and MCP servers) for coding agents.
Platform Engineers responsible for setting up secure, sandboxed environments for AI-driven automation.
Engineering Managers looking to quantify and scale AI-driven productivity across their teams.
-
Modern Dev Stack: High proficiency in Git, CLI environments, and CI/CD pipelines (GitHub Actions, GitLab CI).
Programming Proficiency: Strong command of Python or TypeScript, as these are the primary languages for building Agent extensions.
Containerization: Basic knowledge of Docker (useful for tool sandboxing modules).
Foundation Knowledge: Familiarity with Agentic Architecture (either through our Foundations course or equivalent industry experience).
-
This Agentic Systems deep dive workshop is available for private / custom delivery for your team - as an in-house face-to-face workshop at your location of choice, or as online instructor-led training via MS Teams (or your own preferred platform).
-
The Coding Agent Landscape
Mechanical Deep Dive: Planning loops, scratchpads, and the "why" behind LLM code hallucinations.
Tool Comparison: Strengths and enterprise-fit for Claude Code, Cursor, and GitHub Copilot.
The Extensibility Stack (Claude Code & Beyond)
Project Memory: Using CLAUDE.md to encode standards and architecture decisions.
Building Skills: Creating custom, portable behaviors using the Agent Skills open standard.
MCP for Internal Tools: Building custom MCP servers for secure internal tool integration.
Advanced Implementation Patterns
Subagents & Teams: Foreground/background coordination and handoff patterns.
Event-Driven Automation: Triggering quality gates on tool events.
Context Optimization: Layered prompting, repo summarization, and compaction strategies.
Safety, Security & Quality at Scale
Supply Chain Security: Automated dependency auditing and protecting against package hallucinations.
Testing AI Code: Diff-based generation and measuring "PR Pass Rates" for AI vs. Humans.
Identity & Auth: Just-in-time provisioning for agents and zero-trust principles.
Enterprise Integration & Scaling
CI/CD Workflows: Automated PR reviews and test-failure remediation loops.
FinOps & Governance: Budget caps, API rate controls, and pre-approved tool catalogs.
Trusted by
