Co-Founder and Teacher: An Interview with Jessica Denkelaar

It actually started in Sabrina Ramonov's Skool community, Women Build AI. I started an initiative in there called Women Built Safety, which was about getting a group of women together to build safety apps for women.

I did the sessions together with JoYi Rhyss. She did the mindfulness part and I taught the women how to build. I got a lot of energy from that, and decided to do more sessions outside of the Women Built Safety initiative as well.

Well, first off I taught a group of 60 women how to use the terminal, Git, and Claude Code in the terminal. So funny enough, that is everything you just asked.

I also worked on my vibe code community tool called NoZu (nozu.ai). I had to fix some bugs in the authenticated part of the tool, with Claude Code.

Tools change every week. The thinking underneath them barely changes at all. So I teach the layer that stays.

That means the terminal, Git, and how a project is actually put together. Those have been around for decades and they will outlive whatever launches next month. It also means understanding what the model is really doing when you ask it something: what context it has, what it does not have, and why it will sometimes hand you a confident answer that is wrong.

If what you learned is "click this button in this app", you are back to zero the moment the interface changes. If you learned what that button was doing, a new tool is just new packaging around something you already understand.

The other half is judgment. Knowing what good looks like for your own work, and noticing when the model has drifted away from it. No tool gives you that. It comes from doing the work and paying attention while you do it.

So my measure is not how many tools someone can name. It is whether they can explain what they just built, and why it works.

Many people I teach think ChatGPT is the only tool you can use, and they mostly use it as a Google on steroids. When I tell them about Claude or Google Gemini, I see a lot of aha moments, and a lot of "wow, I want to try this".

I also teach more advanced developer practices to students who have been coding with AI for a while. A lot of what Claude does for them is just not clicking in their head, and when I explain the logic behind it, I get students saying "now I know what I am doing".

That is actually what I always want a student to walk away with. More clarity about what they are doing, not just asking questions to their LLM and not learning anything.

There are many things evolving almost daily. For example, how to make agents work most efficiently right now. That is very much a work in progress.

But since I work with AI and teach it daily, and I have the developer background, I figure it out fast. And then I can teach others what I have learned as well.