Subsurface
NoCode and AI

How I Vibe-Coded an Interactive Fault Seal App with AI

Alan Mourgues
January 8, 2026

Some time ago, through CrowdField, I met Shamim Alam, a petroleum geologist with 35 years of international oil and gas experience and the founder of Geoskillz.

Geoskillz is a free learning platform built for rig-site geologists, mudloggers, and data engineers. It was born from a pattern Shamim repeatedly observed in the field and during interviews: capable professionals held back by inconsistent practical training, limited on-the-job learning, and weak interview preparation.

Instead of academic theory, Geoskills focuses on field-proven, real-world knowledge, delivered in a concise Q&A format. Today, it hosts hundreds of practical questions covering mudlogging, data engineering, and wellsite geology… the kinds of concepts that actually matter on the rig and in technical interviews.

As I’ve been exploring gamified learning, Geoskills struck me as the perfect test case: how do you take solid, text-based educational content and turn it into something visual, interactive, and exploratory?

That’s where the Fault Seal Challenge idea was born.

As usual, the goal here is to show how AI tools can be used to build fully deployed, interactive apps without traditional coding skills.

In this case, I used Google AI Studio, a development environment that can generate complete front-end applications directly from well-crafted prompts (the infamous vibe-coding).

The workflow was deliberately simple and broken into two steps:

  1. I used the Gemini model to brainstorm the game mechanics, refine the learning objectives, and design the overall interaction logic. That same step produced a precise, structured prompt for AI Studio to actually generate the app.
  2. I fed that prompt into Google AI Studio, let it generate the code, and then iterated a few times (minor tweaks, small improvements) until the result felt right.

Once the app is generated, AI Studio outputs a set of files that together form a complete interactive web app.

The next step is deployment… simply making those files accessible on the web so anyone can use the app. If this part is unfamiliar, you can ask Gemini or ChatGPT to walk you through deployment step by step. It’s really not hard to follow.

Once deployed, you end up with a live URL that anyone can interact with.

The final step was embedding the app directly into the CrowdField website.

Here’s the app for you to try yourself:

👉 Launch the Fault Seal Challenge

The YouTube video below shows the app in action (yes, with an AI voiceover):

These tools live under Lab in the top navigation, where I’ll continue publishing more gamified, AI-built learning experiments as they come online.

If you’re an educator interested in interactive teaching tools, or a manager exploring bespoke solutions for your team, feel free to get in touch to discuss potential use cases.

---

PS: If you’d like to try recreating this yourself, I’ve added more details of the vibe-coding process and shared the exact prompts I used in Gemini and Google AI Studio, here.

More Tools

Explore a curated collection of valuable resources in our Store, both free and paid, all designed to help you upskill.

Alan Mourgues is a Petroleum Reservoir Engineering Consultant with 25 years of international experience. He is the founder of CrowdField — the go-to hub for Oil & Gas subsurface professionals to upskill, freelance, and monetize their expertise. CrowdField brings together a global community through: i) Freelance marketplace for niche talent and task-based solutions; ii) Digital Store & Vault of engineering tools, workflows, and resources; iii) AI Hub showcasing startups, workflows, and use cases; iv) Learning resources including webinars, blogs, and curated datasets. Alan’s mission is to empower professionals to turn knowledge into income and future-proof their careers as the energy transition unfolds.

Featured...

All blog posts