From Prompts to Prototypes: Keep Your Vibes Aligned and Focused

AI coding tools have fundamentally changed how ideas can be turned into working digital prototypes. In this hands-on workshop, hosted in collaboration with OST - Ostschweizer Fachhochschule, you will explore a structured approach to “vibe coding,” using AI to rapidly transform concepts into functional prototypes while maintaining clarity, direction, and control.

We move beyond the hype to address the reality of modern development: while AI accelerates coding, it does not replace system design. Participants will learn why structure and planning dramatically improve outcomes and why deployment remains an engineering problem, not a generation problem.

Rather than relying on ad-hoc prompting, you will learn practical techniques for guiding AI through iterative creation, refinement, and troubleshooting. The goal is to build working prototypes while developing a focused and reliable way of collaborating with AI tools, proving that the future is not "no engineers," but better-defined engineering roles.

​💡 Why Join Us?

• ​Build in Real-Time: Gain hands-on experience turning ideas into working prototypes.
• ​Master the AI Stack: Learn how to build software faster using AI tools and discover practical prompting strategies for coding tasks.
• ​Stay in Control: Understand how to stay in control of AI-assisted development.
• Tangible Results: Walk away with a working application you built yourself.

​🛠️ What to Expect

This is a highly interactive, hands-on building session where you will engage in fast-paced prototyping using AI coding tools. Together with the experts from OST, we provide step-by-step guidance on effective prompting techniques and share real-world examples of what works and what fails, ensuring you walk away with a working prototype created during the workshop.

​🎯 Who Should Attend?

• ​AI Practitioners & Builders hungry for better workflows.
• ​Team leaders, Product managers with technical exposure.
• ​Software Engineers and developers eager to master AI-assisted development.
• Students and early-career engineers.
• Anyone interested in building software with AI tools / tinkerers.

💻 What to bring

• A laptop
• Docker installed
• Github co-pilot or Claude code
• Integrated Development Environments like Zed, VSCode, zed/vscode/intellij
• Copilot or claudecode credits (free tokens may not be enough)
• If you have lovable, replit etc., you are welcome. We can compare the outputs.

🗓️ Our Agenda

The core of this session is understanding the critical gap between a working demo and a production system. Through hands-on exercises, we move beyond simple generation into real-world engineering.

Phase 1 | Setup: Development Environments & DevContainers

We begin by introducing a standardized development setup using devcontainers.

Phase 2 | Exercise 1: The Vague Prompt

We start with a minimal, underspecified prompt to generate an AI solution.

Phase 3 | Exercise 2: Plan First, Then Prompt

Lastly, we shift gears by writing a structured plan before using AI for implementation.

Note on Participation: This first session is completely free of charge. Subsequent sessions in the series will be available for a registration fee. Please note that as space is limited to ensure a high-quality hands-on experience, we reserve the right to select participants.