This week, I had the honor of presenting at the Alamo STEM Ecosystem meeting, where I was invited to talk about artificial intelligence and its impact on education. As an educational consultant with years of experience in bilingual and ESL education, instructional design, and the integration of educational technology, my goal was to showcase the work I’ve been doing over the past two years with AI in teacher education, and how these tools can serve as a bridge between the professional knowledge teachers need and the complex realities of today’s multilingual classrooms.

The central focus of my presentation was sharing the story behind Verónica, the AI-powered tutor I’ve been developing. Verónica isn’t just a chatbot. She’s a fully integrated system on our website, designed to support both pre-service and in-service teachers with real classroom needs. Her core purpose is to enhance teacher training by generating scenario-based learning activities, offering expert support and best practices for teaching academic language in subjects like science in multilingual settings, and helping educators develop and practice academic vocabulary in the target language (particularly Spanish) for those preparing for bilingual certification. Verónica exemplifies how AI can amplify our capacity to train and empower teachers, not replace them.

However, perhaps the most important message I shared came at the end: you don’t need to be a professional coder to shape the future of education. This is where the concept of vibe coding comes in.

Coined by computer scientist Andrej Karpathy in February 2025. Vibe coding refers to a development approach where individuals use natural language prompts to instruct AI models to generate code, shifting the programmer’s role from manual coding to guiding and refining AI-generated outputs. Karpathy described this method as a process where he would “see things, say things, run things, and copy-paste things,” emphasizing the conversational nature of interacting with AI to produce functional software. It’s a movement embraced not only by non-coders but also by experienced developers and engineers. These tools help us translate our ideas, especially those informed by classroom realities, into functional, real-world solutions.

Over the past year, I’ve embraced this idea of vibe coding in my own work, so much so that I even wrote a LinkedIn article about it. While I have a background in coding through my master’s degree in Computer Education and Cognitive Systems, what excites me most is how accessible these tools have become. I’ve implemented some of these Vibe Coding projects with my own kids (14 and 11 years old), and it’s incredible to see what they can build, even with no formal coding experience. Vibe coding allows anyone to bring ideas to life, but for those of us with a technical foundation, it dramatically boosts productivity. In my case, I’ve used this approach to develop a wide range of new features for Verónica and our scenario-based learning system. Most of these tools are built using a standard web development stack: HTML and JavaScript for the front end, PHP for the back end, and SQL for managing databases. It’s a foundational skill set for web development, and now, with AI support, it’s within reach for educators, creators, and students alike.

I believe that if educators, especially those of us who understand the complexities of multilingual and multicultural classrooms, get involved in learning how to use these tools, we can build the future of education ourselves. We can create mobile apps, web-based tools, and customized learning platforms for learners in any language. Large language models today are multilingual by design, giving us the opportunity to create inclusive solutions that honor the diversity of our classrooms. These new AI tools can empower minority communities by lowering the barrier to creation and amplifying voices that have long been underrepresented in the tech world.

My journey from bilingual education to AI integration has taught me that educators have something unique to offer in the AI revolution: insight, empathy, and experience.

Empowering Educators and Students Through AI

One of the most exciting aspects of integrating AI into education is its potential to empower both teachers and students, not just as users of technology, but as creators. With the support of AI and accessible tools like HTML, JavaScript, and no-code platforms, educators can design customized classroom solutions without needing to become full-time programmers.

For teachers, this means the ability to build interactive games, assessment tools, or lesson generators tailored to their students’ needs. It also opens the door to automating simple, non-sensitive tasks (like generating activity templates, lesson planning tools, or self-check quizzes) that support instruction and engagement without handling student data or replacing the thoughtful, personalized feedback only a teacher can provide. Even simple projects, like a JavaScript-based game I developed to help students practice accent marks in Spanish, can make learning more engaging and effective. You’ll find the game embedded at the bottom of the blog post “La acentuacion su importancia y reglas”.

Beyond technical utility, learning to build with AI models and basic code allows educators to model lifelong learning and creativity for their students. It positions them to support student interests in coding and digital problem-solving, even if they aren’t teaching programming directly. In multilingual classrooms, AI-driven tools can be especially designed to meet students where they are, making content more accessible and interactive across languages and cultural backgrounds.

By embracing these tools, educators not only increase their own productivity but also unlock new pathways for student engagement, collaboration, and inclusive learning design.

Implementing AI Responsibly

As educators and creators embrace vibe coding and begin building tools using AI, it is essential to remember that responsibility and security must guide innovation. While the ability to generate working code with the help of AI is empowering, it also comes with the responsibility to follow best practices, particularly when handling student data (which I strongly advise against unless you’re working within a secure, FERPA-compliant system and have the appropriate permissions in place) or deploying tools in online environments. Secure platforms, thoughtful architecture, and proper handling of API keys and user input are critical to prevent vulnerabilities or breaches that could compromise trust.

In developing Verónica, I prioritized a secure and robust implementation strategy:

• Secure Platform: Our website serves as a protected environment, ensuring that user interactions—especially from educators and learners—are safeguarded.
• API Integration: Inputs from interactive forms are transmitted via secure API connections to OpenAI’s models, allowing real-time feedback without exposing sensitive data.
• Technological Stack: The system runs on a tried-and-true foundation (PHP for backend processes, HTML and JavaScript for frontend interaction), offering both stability and flexibility.

By adhering to best practices in API key management, code validation, and data protection, we’ve created a system that not only leverages the power of AI but also maintains the integrity, safety, and privacy that educators and students deserve.

Learn for Yourself, Teach Your Students

As I said before, coding isn’t just for computer science specialists anymore. As a teacher, learning basic coding skills can transform both your professional development and your classroom. With AI tools like Claude, the learning curve is gentler than ever before, making this the perfect time to dive in, whether to create your own teaching tools or introduce students to this essential 21st-century skill.

The Minimal Setup

Whether you’re learning for yourself or preparing to teach students, start with these basics:

• A computer with internet access
• A modern web browser
• Access to a Large Language Model such as ChatGPT, Claude, or Gemini (I recommend Claude with free access via https://claude.ai, more about it below)
• A beginner-friendly code editor (some recommendations below) 

Which Programming Language Should You Choose?

I get this question from fellow teachers all the time, and here’s my straightforward advice:

For your own professional tools: Start with Python. Trust me on this one. The syntax reads almost like English, which means you’ll spend less time fighting with cryptic symbols and more time creating useful tools. Python is forgiving for beginners but powerful enough to grow with you.

For web-based projects: If you’re dreaming of creating a class website or interactive web tools, you’ll eventually need to learn HTML, CSS, and JavaScript. HTML structures your content, CSS controls how it looks, and JavaScript adds interactivity. While JavaScript can have a steeper learning curve than Python, HTML and CSS are more accessible and provide a solid entry point into web development. Many educational platforms and student-centered projects rely on this trio of technologies to bring ideas to life on the web.

For teaching students:

• For elementary/middle school: Start with Scratch to introduce programming concepts without the barrier of syntax.
• For middle/high school beginners: Python offers the most accessible introduction to text-based coding.
• For students interested in web development: Begin with HTML, then add CSS, and finally JavaScript, in that order.

You don’t have to master everything at once. While you can use a variety of AI assistants (such as ChatGPT, Gemini, or others) to support your coding journey, I ultimately chose Claude as our primary coding companion. In my experience, Claude excels at explaining programming concepts with clarity, offering beginner-friendly examples, and maintaining a natural, conversational tone that makes learning to code feel less intimidating. Whether you’re troubleshooting an error, trying to understand a new concept, or building something from scratch, Claude can guide you step by step in any of the languages mentioned.

Teacher-Friendly Coding Environments

These platforms work equally well for your own learning or classroom use:

1. Replit (replit.com)
   • Web-based environment requiring zero setup
   • Supports multiple languages (Python, JavaScript, HTML/CSS)
   • Built-in sharing features for classroom collaboration
   • Create private and public projects
   • Classroom management features available for educators
2. CodePen (codepen.io)
   • Perfect for web development projects
   • Shows code and visual results side-by-side
   • Great for creating interactive teaching materials
   • Easy to share examples with students
3. Scratch (scratch.mit.edu)
   • Block-based visual programming
   • Excellent for beginners and younger students (8-16)
   • Helps build logical thinking before syntax concerns
   • Robust teacher resources and classroom management tools

How to Use Claude as Your Personal Coding Tutor

For Your Own Learning:

• • Start with practical classroom tools – Ask Claude: “Help me create a simple grade calculator in JavaScript”, “How do I build a random student picker in Python?”
  • Learn at your own pace – Ask for explanations: “Can you explain how variables work in Python?”
  • Request simpler analogies – “Explain functions using teaching concepts”
  • Build progressively – Begin with small, useful projects relevant to your teaching. Ask Claude: “What’s the next step after I’ve learned basic Python syntax?”
  • Troubleshoot with confidence – When your code doesn’t work, copy the error and ask Claude
  • Request explanations – “Why isn’t this code working? Here’s the error...”

For Teaching Students:

• Scaffold learning experiences – Ask Claude for age-appropriate coding projects. Example: “What are good first Python projects for 7th graders?”

Differentiate instruction Get help creating tiered coding challenges for different ability levels. Example: “How can I modify this code to make it easier/harder?”

Prepare explanations Have Claude break down concepts in student-friendly language. Example: “Explain how loops work using a classroom analogy.”

Create multilingual coding activities Ask Claude to generate project instructions or comments in Spanish, or any target language, so students develop both coding and language skills. Example: “Create a Python project for middle school students with instructions and comments in Spanish.”

Reinforce academic vocabulary through code Have Claude design activities where students must include subject-specific terms in their code (e.g., math, science, social studies). Example: “Design a Scratch or Python project where students create a quiz about ecosystems using science vocabulary.”

Promote language use through creative outputs Ask for coding projects that require students to display definitions, write explanations, or narrate steps as part of the final product. Example: “Create a coding project where students build a calculator that explains each math operation in Spanish.”

Step-by-Step Example: Creating a Classroom Tool

Here’s how to create your first safe and fun classroom tool without using actual student names:

1. Go to replit.com and create a free account
2. Start a new Python project
3. Open Claude in another tab and ask Claude: “I’m a teacher learning to code. I want to create a simple random selector tool in Python that I can use during class activities. Instead of using student names, I’d like to assign each student a fun animal nickname (like Dolphin, Panda, Tiger, etc.). The tool should let me: Enter a list of animal names, and randomly select one, and mark that name as ‘called’ so it won’t be repeated until all have had a turn”
4. Paste the code Claude gives you into Replit
5. Test it out! Try using a list like [“Tiger”, “Koala”, “Penguin”, “Fox”, “Giraffe”]
6. Ask Claude follow-up questions like: “How can I make this list reshuffle once everyone’s had a turn?”, “Can I display a fun message with each selection?”
7. Ask Claude to explain any parts you don’t understand
8. Gradually customize it to better fit your classroom needs

Step-by-Step Example: Teaching Students to Code

When introducing students to coding:

1. Create a free teacher account on Replit
2. Set up a class and invite students
3. In Claude, ask: I’m teaching coding to 5th graders for the first time. We want to create a simple rock-paper-scissors game in Python that they can implement in Replit. Please provide the simplest version with explanations for each part that 10-11 year olds would understand. Include some simple challenges they can try after getting the basic game working.
4. Use Claude’s response to create a starter template for students
5. Share the template with students through Replit
6. Guide students through understanding and modifying the code
7. Use the extension challenges Claude suggested for early finishers

Safety and Best Practices

When learning yourself:

• Start with small, manageable projects related to your teaching
• Commit to regular practice (even 15 minutes daily helps)
• Join teacher coding communities for support
•  

When teaching students:

• Review all AI-generated code before sharing with students
• Start with offline/local projects for younger students
• Teach responsible AI use – Claude helps learning, not replacing thinking
• Consider your school’s technology policies regarding AI tools
•  

Learning and Teaching with a More Knowledgeable Other

Here’s what makes working with AI truly revolutionary: you can just talk to it! That’s the beauty of large language models. You can use the audio feature to speak your ideas, questions, or coding problems aloud, just like you’d talk to a trusted colleague.

When you’re stuck on a coding problem:

• Talk through what you’re trying to do
• Paste the code you’re struggling with
• Explain what you expected to happen
• Share any error messages
• Ask for explanations in plain language

After a while, you’ll start to feel its rhythm and “vibe”. It becomes less like formal programming and more like a collaborative conversation, where ideas flow freely and support feels immediate. That’s what makes these tools so powerful. We can express our classroom needs in our own words, not in technical jargon.

One thing I strongly recommend (especially if you really want to learn to code) is to take the time to read the code you’re getting. Try to understand what each part is doing. When you know what you’re looking at, you can ask better follow-up questions and request explanations for specific sections or functions. That’s where real learning happens.

While it’s likely that in the future many of us will rely on AI to do the heavy lifting in coding, it’s still critical to understand the mechanics and logic behind the code. The real value is in the thinking, the problem-solving, and the reasoning that go into each step.

I’ve found myself in frustrating loops where the AI gives me broken code, and I keep getting new versions that also don’t work. But when I finally stop and read through the code carefully, I often realize the issue is in the AI’s response, with some mistake in naming or referencing. I point it out, and the AI will “apologize”—in quotes, of course, because it’s not really thinking. Still, that moment is where the real learning happens.

This process, of questioning, reading, adjusting, and understanding, is a powerful way to learn. If you engage with it fully, you’re not just building tools. You’re building your own skill set and confidence, one interaction at a time.

Unlike static tutorials or traditional programming resources, AI assistants adapt to your specific situation, your teaching context, and your learning style. In many ways, they take on the role of what Lev Vygotsky called the more knowledgeable other—someone who helps guide you through the zone of proximal development (ZPD). AI doesn’t just hand you answers—it provides just enough scaffolding to help you understand and move forward. Every coding challenge becomes a learning opportunity, where the AI helps you troubleshoot, rethink, and solve problems at the edge of your current ability. It’s a dynamic partnership that supports growth, mirrors natural learning, and keeps the momentum going.

Learning to code as a teacher (whether for your own use or to teach students) doesn’t require extensive training or a technical background. With AI tools like Claude or ChatGPT acting as your patient, always-available tutor, you can progress at your own pace, build tools that make your teaching more efficient, and introduce your students to skills that will serve them well into the future. And the beauty of these large language models is that they are multilingual, so they can help you teach your students in any language, allowing them to practice in their target language with meaningful, personalized support.

The best part? We teachers are the ones who truly understand the challenges of the classroom and the needs of our students. When we create our own tools, we’re building the future of education from the inside out. Nobody else has our perspective or our dedication to improving learning.

By embarking on this journey, you’re modeling exactly what we hope to instill in our students: curiosity, problem-solving, and lifelong learning. And that might be the most valuable lesson of all.

Let’s not wait for someone else to build the future. We can build it ourselves!