Why this workshop?
AI coding assistants are evolving at a pace that is hard to keep up with.
New models, tools, and workflows emerge every week and most developers are using only a fraction of what these tools can actually do. There is a big gap between casually chatting with an AI and truly mastering agentic, production grade workflows.
This workshop bridges that gap. It is built on real experience, not tutorials, including lessons earned the hard way from building production apps with AI coding assistants, including the pitfalls of vibe coding and how to recover from them.
What will you walk away with
What Will You Walk Away With?
A deep understanding of how AI coding assistants work under the hood: models, memory, context, tokenization, and tool use
Practical prompting techniques and agentic workflow patterns you can apply immediately
Direct experience building hooks, custom commands, skills, and sub-agents with Claude Code
A proven spec driven development workflow (Spec to Code, Code to Spec)
Best practices to avoid the common traps of AI assisted development
Live demos using DevoxxGenie, an open source JetBrains plugin with 45K+ downloads, showing how these concepts apply directly inside IntelliJ IDEA
The confidence to evaluate, adopt, and extend AI tools that genuinely amplify your productivity
Who is it for?
software developers, tech leads, and engineers who want to go beyond basic autocomplete and understand how to harness AI coding assistants at a professional level
Whether you are already using Cursor, GitHub Copilot, or Claude and want to understand the deeper mechanics, or you are evaluating how AI fits into your team's workflow, this is for you
IntelliJ IDEA users will get particular value from the DevoxxGenie (https://genie.devoxx.com) demos, seeing the concepts come to life directly inside their IDE
No prior AI or ML background required
Any programming language background works
Workshop breakdown
Morning (09:00–13:00) - Foundations
Covers the 10,000 foot view of the AI coding landscape, how LLMs work, the role of prompts, memory, context windows, tools, MCP (Model Context Protocol), and agents.
Includes a walkthrough of Anthropic's ecosystem and Claude Code CLI, with live demos inside IntelliJ IDEA using DevoxxGenie.
Afternoon (14:00–18:00) - Deep Dive & Hands-on
Practical, interactive sessions covering Hooks, Custom Commands and Skills, Sub-agent architecture, PSI and LSP integration, Plugin systems, Spec to
Code workflows, Agent loops, and Best Practices
Each section includes a live demo or hands on exercise, many of which are demonstrated using
DevoxxGenie to show how these patterns work inside a real, actively used tool.
Who is the trainer?
STEPHAN JANSSEN
Founder
Devoxx
Stephan Janssen is a veteran of the Java and developer community, with over 25 years of experience building communities, conferences, and software.
He founded the Belgian Java User Group (BeJUG) in 1997, launched JavaPolis (now Devoxx) in 2001, which has grown into a global conference family welcoming 20,000+ attendees per year across Belgium, the UK, France, Poland, Morocco, and Greece. He also founded VoxxedDays in 2014, expanding the community further into Switzerland, Luxembourg, the Netherlands, Italy, Romania, and beyond.
Stephan is also the creator of DevoxxGenie (https://genie.devoxx.com), an open source JetBrains plugin with 45K+ downloads that brings AI coding assistants directly into
IntelliJ IDEA. The plugin supports multiple LLMs, agent mode with tool execution, and an MCP server marketplace, and is actively used by developers around the world who want AI assistance without leaving their IDE.
Today Stephan applies that same community first mindset to the rapidly evolving world of AI coding assistants, sharing practical, production tested insights with developers navigating the shift.
See also https://stephanjanssen.be
