Blog - Sunday 28 December

Employee Spotlight: Julian von der Goltz

Meet Julian von der Goltz, ML Engineer at KickstartAI. With a background in aerospace engineering and computer visios, he brings a strong technical foundation to the team. His work focuses on developing AI solutions that help partner organizations adopt new technologies while creating positive societal impact.

Let's dive into his journey, what drives his work, and how he aims to support the Dutch AI ecosystem.

What’s your role at KickstartAI?

I’m a machine learning engineer. In short, my role is about developing cutting-edge AI solutions for our partner companies and other organizations, and helping them adopt these technologies effectively.

Our AI projects vary a lot, but what connects them is the goal of using the technology to tackle problems that have a positive social impact.

For example, I’m currently involved in our project with the Leiden University Medical Center (LUMC), where we’re developing an AI tool to improve wound care and reduce nurses’ administrative workload. I’m also part of our project with ING, where we’re working together to build a chatbot for financial health.

What’s your background, and how did you enter the AI industry?

I started out studying aerospace engineering and I properly got into AI during my masters, when I focused a bit further on robotics. Naturally, part of robotics is AI and computer vision, and that’s when I started to dive into machine learning and deep learning. After graduating, I spent about six years as a computer vision and software engineer at a startup that builds drone inspection systems for airplanes. From there, I joined KickstartAI to apply my experience to new domains.

What do you enjoy about being part of KickstartAI?

Our mission for societal impact is definitely one thing. When I first applied for the job, I underestimated how rewarding that would be. Our work isn’t just about maximizing revenue or improving the usual KPIs, but also about helping people. I really like that. I also enjoy working together with a super smart, and super motivated team, where there's plenty of room to explore and try out different technical ideas/technologies.

How does your work impact the AI industry in the Netherlands?

I think the most effective part of my work here is taking part in the challenges we do together with our partner companies. Imagine these challenges as structured, two-week hackathons, where members from all the different organizations come together to start developing an AI solution for a real-world problem.

You meet a lot of people and work intensively with them. Everyone contributes their experience and then you all leave with shared insights and knowledge. I think this has a positive impact on the Dutch AI ecosystem. The challenges might be limited to our partner companies, but we always share the learnings on our blog.

What drives you in your work?

A big driver for me is my curiosity about technology and the excitement of solving problems with it. Even before the big AI boom, I loved how software engineering let you build almost anything with code.

Now, with AI, you can create solutions faster, smarter, and on a completely new level. I’m working on things today that I wouldn’t have had the time or tools for just a few months ago. That’s very cool/motivating!

What are your go-to AI tools?

I mainly use coding assistants, such as Codex, Droid, and Gemini. I see these tools as productivity enablers, although you still have to be careful when using them. Sometimes they make mistakes, and you can easily end up spending more time fixing something than if you had done it yourself. So I try to be mindful about when to use them and when to rely on my own code.

I often switch between different coding assistants, depending on the task. I prefer those that let you change the underlying model without any reprogramming. Each model has its own strengths and weaknesses, and I like using the best tool for each job.

What’s your advice to people considering a career in AI?

The classic advice is “learn the math.” But I’d like to offer another perspective: try to understand the processes behind the AI models. How are these models made and trained? Why do they produce certain outputs? Why do they behave in a specific way?

That understanding helps a lot when you start building with or on top of AI yourself. You’ll know what to expect and what its limits are.