A modular, hands-on course on agentic software engineering. It runs to roughly twelve hours of lectures and guided practice, and can be delivered in full or as standalone modules tailored to your team. Each module pairs short lectures with practical exercises on real workflows.

Context

Why still program, and what does it mean now? We frame what agentic programming actually is, where it succeeds and where it fails, and the different styles in use today — from micro-editing and specifying to stress-testing, reviewing, translating, and fully orchestrating. (~0.5h lecture)

General Techniques

The core skills: how to prompt effectively, how to assemble the right context, how to orchestrate agents toward a long-term goal, and how to make them autonomous. (~0.5h lecture · ~0.5h practice)

Human in the Loop

Where people stay essential — what can be delegated to AI and what cannot.

  • Architecture — should architecture stay human-owned, and how can AI help design it? (~0.5h lecture)
  • Testing — is testing more important than ever, and can AI generate it? (~0.5h lecture · ~0.5h practice)
  • Security — is “vibe coding” insecure, and how do you use AI to harden your code? (~0.5h lecture · ~0.5h practice)
  • Reviewing — how to code review with agents, the role of static analysis, and whether documentation still matters. (~0.5h lecture)
  • Integration — plugging AI into continuous integration, and controlling your agents remotely. (~0.5h lecture · ~0.5h practice)

New Software Engineering

How the discipline itself changes: which old roles and methods survive, how a software organisation is structured in the AI era, how to compose a good team, where productivity gains really come from, and how to price software now. (~1h lecture)

Deployment

Making it real and sustainable: concrete workflows, monitoring agent performance, driving down cost, staying sovereign and provider-independent, the role of local models, and choosing the right model and ecosystem for the job — with benchmarks. (~0.5h lecture · ~1.5h practice)

Hands-on Techniques

The practical toolkit: MCP, RAGs, tools, bugbots, skills and rules, and modern agent harnesses. (~0.5h lecture · ~1h practice)

Per Use Case

What solutions fit when — worked through concrete examples: self-improving websites, porting code, scientific software, prototyping, GUIs, and more. (~0.5h lecture · ~2.5h practice)

Interested in a session for your team? Get in touch.