AI on Rails: Level Up the Way You Code with AI

Ruby developers are learning to generate code with AI. But are we learning how to work with it well?
About this Event

On August 4, AI Tinkerers Paris, Paris Ruby workshop and ParisRB.new are joining forces for a hands-on evening at Algolia.
You will discover how experienced builders use AI in real development workflows, question them about their tools and methods, and then put those practices to the test on a Ruby kata.
The goal is not to find out which AI writes code fastest.
It is to improve the new skills that matter when coding with AI:

• giving an agent the right context;
• breaking a problem into useful tasks;
• choosing between one agent and several;
• reviewing and testing generated code;
• identifying plausible but incorrect solutions;
• knowing when to guide the AI—and when to take back control.

You can work individually or with other participants. Preparing or presenting a demo is entirely optional.

Learn from real AI coding workflows

The evening will begin with short technical talks from builders using AI in their development work.
Beyond the projects themselves, we encourage you to ask practical questions:

• Which tools and model versions do you use?
• How do you structure prompts and project instructions?
• When do you use one agent or several?
• What context do you give the AI before it starts coding?
• How do you review and test generated code?
• Which tasks do you still prefer to handle yourself?
• What mistakes or failure modes do you encounter most often?

No high-level pitches. Expect working code, concrete methods and honest lessons from real projects.
Participants who prefer to start building immediately may begin the workshop during the talks. The speakers will remain available afterwards, so you can continue the conversation and ask for practical guidance while working on your kata.

Put those practices to the test

Choose a Ruby kata suited to your level and solve it using your preferred AI coding setup.
You might use:

• one coding agent from start to finish;
• several agents for different tasks;
• an AI-assisted IDE;
• a terminal-based agent;
• your own prompts, rules or development framework.

The katas range from beginner Ruby exercises to Rails APIs, Action Cable, testing and production-level challenges.
Use the workshop to try one of the practices shared by the speakers—or compare it with your own method. Speakers and other experienced builders will be available to discuss tools, prompts, context management, agent workflows, testing and debugging approaches.
The result matters, but the method matters more.

Turn your experiment into a useful demo

At the end of the workshop, you will have time to turn what you learned into a short technical demo.
Two AI coaches will be available:

• Use our AI Tinkerers technical feedback agent to review the technical content of your presentation: the problem, approach, architecture, tools, results and lessons.
• Use Speakli to practice your oral delivery and receive feedback on how clearly and concisely you explain your work.

Your demo can cover:

• the kata you selected;
• the coding tool and model version you used;
• the context and instructions you gave the AI;
• whether you worked with one agent or several;
• a practice from a speaker that you tested or adapted;
• what worked—and what failed;
• how you tested and corrected the generated code;
• one practice other builders could reuse.

Preparing and presenting a demo is entirely optional.

Share your best practices

The most instructive demos will be invited to present their methods on stage.
We are not only looking for completed solutions. A strong demo might reveal:

• an effective way to provide context;
• a useful multi-agent workflow;
• a rigorous method for testing generated code;
• an unexpected failure mode;
• a smart human intervention;
• a lesson that could save other builders time.

If your demo is not presented on stage—or if you prefer not to speak publicly—you can still share your setup, results and lessons directly with other participants during the networking session.
The goal is not to turn every participant into a speaker. It is to make the best practices discovered during the workshop available to the whole community.
The most instructive demos will receive goodies.

Agenda

🕑: 07:00 PM
Welcome
🕑: 07:15 PM
Introduction

Info: Meetup, community and sponsor introductions, followed by a presentation of the available katas.

🕑: 07:30 PM
AI Coding Practices: Lightning Talks and Q&A

Info: Learn how experienced builders use AI in their development workflows. Ask about their tools, model choices, prompts, agent workflows, testing methods and common failure modes. Participants who prefer to start working may begin their kata.

🕑: 08:15 PM
Put the Practices to the Test: Ruby + AI Workshop

Info: Work individually or in a small group. Test one of the speakers’ practices or compare it with your own. Pizza, soft drinks and beer will be served during the workshop.

🕑: 08:45 PM
Demo Preparation and AI Feedback

Info: Optionally prepare a short demo. Use our technical feedback agent to strengthen its content and Speakli to practice your oral delivery.

🕑: 09:00 PM
Selected Demos: What Worked, What Failed, What We Learned

Info: The most instructive demos will share their tools, methods, results and lessons on stage.

🕑: 09:30 PM - 10:00 PM
Networking

Info: Continue the conversation, explore the other experiments and exchange best practices directly with their builders.