Learning While Earning: Three Months Inside Studio JDB with Owethu

Our intern, Owethu, joined Studio JDB at the start of April this year, fresh out of university and fluent in Revit. Three months later: a live commercial project shipped, AI taught to model on command, and a 400-year-old castle loaded inside SketchUp. Joshua sat down with Owethu to talk about what the internship actually looked like, condensed here from their full conversation.

Arrival: expectations vs. reality

Joshua: Take us back to your first week. What did you expect the internship to look like, and what was it actually like?

Owethu: From the get-go, there was a lot of work, which was the weird thing for me compared to past experiences. When I got here, it was more of learning while earning; it was benefiting me while I was growing as a person. It wasn't what I expected, even though I'd researched the studio beforehand.

"Not setting high expectations was what benefited me. It was a supportive environment - it helped me learn and try to prove myself, just for the fact that you allowed patience with me while I was learning new software."

You've described yourself as more of a software person than a design person. When did you realise that?

While I was studying. I really enjoyed 3D software from the get-go, more than the design part. I learn software faster than design principles. Here in the studio, I realised I liked the AI and modelling side more than sitting and designing. I started the internship doing a little bit of everything, but ended up on the AI and experimenting side, and Joshua supported me in pushing and learning more of that.

What were you trusted with before you felt ready?

The commercial project, a store-within-a-store for cellular devices, where the client wanted every device on the stand modelled as the actual device. There was a lot of pressure to meet the deadline, and I was learning as I went, picking up my mistakes as I made them. I wasn't ready to take that on, but I was trusted to take it on, which was pretty positive. And Tenneal was supervising, so there was a safety net, even though the work I did was the work that got submitted.

Trust before readiness - with a safety net.

Interns do real work that gets submitted, but a senior designer carries final responsibility. That combination is what accelerates growth.

University vs. the real studio

Were there tools or ways of working that university didn't prepare you for?

We were mostly taught the longer, older ways — rather than being taught to set up templates properly, which makes everything way faster. You have to learn that on the job, assisted by the people already working there.

"Asking questions really goes a long way. Trying to figure something out yourself might take 30 minutes, when asking could solve it in a few seconds."

And in a five-person studio, is there a balance between asking a person and asking an AI?

Yes, it's what I do on my own, too. In a studio of five people, there are only so many people you can keep bothering. Sometimes it's just going to be you. But I've always been given the space to ask.

How does your final university project compare to a real project?

A mistake in school only affects you and your marks. In a real studio, it affects real people and real environments. In school, you do everything properly and accurately because the lecturer inspects your whole file. In a real studio, you're producing what fits best for the client while saving time — the client won't go around inspecting your model, because the client doesn't understand the software.

School optimises for accuracy; practice optimises for the client and the clock.

Both matter, Joshua notes that skipped fundamentals (like unassigned tags and layers) resurface as real delays when a manufacturer's laser cutter reads your file.

Living in two worlds: SketchUp and Revit

You came from Revit. What felt strange about SketchUp — and what clicked?

The surprise factor was how free SketchUp is. You can model literally anything without having to know the physics of it, even as a beginner. Revit is accurate and rigid — a curvy building takes extensions, automations, and deeper knowledge of the software. SketchUp was easy to learn, and for me, it has more room for growth.

And Revit's biggest advantages?

It almost mimics your real-life environment. You can collaborate with other professionals in the same software without converting files, and your sheets and everything live in one place — something I hope SketchUp ends up having.

Does BIM thinking have a place in a design-led tool like SketchUp?

I recently started learning BIM properly, working from home. I used to see it as this scary, different field with complex terminology - your BEPs, your LODs. But when you get the explanations, it's simple things you already do, structured properly. And BIM is not necessarily known as Revit - BIM is a process on its own. It can be applied to SketchUp, too.

"BIM looked like a scary field of its own — but it's simple things you already do, structured properly. BIM is not know Revit. BIM is a process on its own."

AI in the workflow

You spent your last month researching AI - having SketchUp draw from a text prompt via Ruby scripts. What stuck?

Understanding a bit of the coding language, so I know the patterns — if you structure a script one way, it works out how you want, compared to another way. I documented that for myself. It helps in the early stages of design so you don't spend two hours on something that wouldn't work out. The same goes for white-boxing with Mbongeni: you figure out if something works before you model it up and waste time.

What didn't work?

I didn't get frustrated — one day it will work the way I want. But I didn't save my scripts. The next day I'd want the same counter at a different size, ask the AI with the same prompt, and get something different. Prompt variations differed quite a bit, and that affected the outcome.

Save your scripts.

AI output isn't repeatable from the same prompt. Treat working scripts like assets, and expect to maintain them as software versions change, the way coders do.

Is AI an advantage, a moving target, or there to steal your job?

If you use it properly, it's an advantage. On projects, it fastens the workflow a lot, instead of placing every component individually to see if it fits, AI cuts down the time and tells you what won't work before you spend two hours modelling.

What would you say to people who feel threatened by it?

Just try it out. Being open to trying new things really helps. I could have been stubborn and said, "I'm a Revit user, I won't use SketchUp" - but here I am, because I tried something and it worked out pretty well. We should have the same mentality with AI, because it's only going to keep improving.

"Catching the wave now will help a lot going into the future."

Reality capture from a junior's seat

Context: In December 2025, the studio LiDAR-scanned the Castle of Good Hope for a Jameson distillery tour activation - there's no DWG file for a castle from the 1600s. During Owethu's internship, the XGRIDS K2 scanner launched, and Owethu learned to work with Gaussian splats and point clouds inside SketchUp.

What does having a captured site change in your day-to-day drawing?

It helped a lot - it was a surprise factor. I came from a modelling context of maps and street view, trying to match things to scale just to give context to what I'm designing. Having a way to capture everything and put it straight into your software, and just have it there, that was really exciting.

What went through your head the first time you saw a scan load?

Wow. It was a really big eye-opener. We have this technology, and firms are still modelling everything out by hand. Going into the future this will be more relevant — it's good that we're experimenting now, while it's still at the beginning phase.

You don't need the expensive scanner to start.

Apps like Scaniverse let you practice reality capture on your phone - at home, in a restaurant, anywhere. The skill transfers when the professional hardware arrives.

Advice and next steps

How did you find your branch: your specialism? And how should students find theirs?

Being open to trying new things, having the curiosity to try it out — you never know what might fit you. Don't only use software when a school brief forces you to; spend your own time experimenting at your own pace. Experimenting and curiosity are the main drivers.

Some would say you've gone from designer to architectural technologist. What about students who feel that's giving up on being an architect?

There are a lot of alternatives in architecture — it's not only conceptual design or only technology. Experiment and you'll eventually find what's meant for you. And the reality is: you can only have so many designers.

"We have a lot of design-driven people and not so many software-minded people. The 'how-to' people are scarcer."

What are you deliberately nurturing right now?

I set aside two hours of my day to learn software — the software itself, not a project. Going forward, I want to be diverse: not just Revit and SketchUp, but tools like Meshy AI and Hyper3D, where you can make a 3D object from an image. It comes down to keeping the mindset to learn and carry on learning.

Last one: what should we do better for the next intern?

The briefing. Coming from school where the brief is detailed, arriving here to a four-liner was difficult. And the library, everything was almost everywhere.

Joshua's note: fair, and taken. We're thin on standard operating procedures, and we're fixing that, building out our knowledge base so the next intern can open it and understand not just the steps, but the way of thinking behind them.

Thank you, Owethu, for your time, your work, and your candour. We're watching your journey between Revit and SketchUp with interest.