Neus, Team Lead Data Science

That’s a good question I’ve been asked often. I actually started at CHECK24 as a Product Manager, even though my background as a physicist probably didn’t fit that position all that well. At the time, I was a postdoc and wanted to move into industry to apply my skills to practical use cases. A colleague and friend told me about an opening, and I was excited by the challenge. Product Manager? At CHECK24? Sounds interesting! Shortly afterwards, I joined through the employee referral program as an API Product Manager.

Looking back, it was the perfect way to combine learning the product itself with a role that carried managerial responsibilities. I learned a lot about product managers’ daily work and their internal processes and customer‑related decisions: how they work hand in hand with developers; how to prioritize, plan, and design a project; and the tools used to deliver on time. I was also lucky to have great mentors.

After a few months, I gradually began taking on projects that were more Data Science‑related and worked on them outside the API team, such as recommendation systems and customer segmentation. I wrote the code, validated the algorithms, and implemented the changes myself—until my tasks no longer matched those of a Product Manager. It was a seamless transition that happened quite naturally, so I sat down with our managing director and we agreed it was time to adjust my job title to the more appropriate “Data Scientist.” The time I spent as a PM has really helped me gain a better overview of the entire product, which I now see as an advantage.

I mainly work with Python for Data Science, Machine Learning, and rapid prototyping, frequently using libraries such as pandas, NumPy, scikit‑learn, PyTorch, and TensorFlow. For querying data and optimizing performance, I use SQL, particularly with relational databases like MySQL and PostgreSQL. On the infrastructure side, I work extensively with AWS—for example, S3, Glue, and Athena. We train and deploy models with MLflow; visualizations and dashboards are built in Tableau; and for version control we use Git with Bitbucket.

This is probably one of the best things about working here: you never stop learning. On the one hand, there’s an extensive training catalog and a learning & development budget I can use flexibly—whether for online courses, books, or in‑person trainings. On the other hand, I especially appreciate the “share knowledge” motto: through internal tech and AI conferences, the cross‑vertical Data Science group, and regular exchanges about tools and projects, we share knowledge with one another. I find it very valuable that there are always experts in the community to reach out to and that colleagues present new technologies in tech talks. This creates a strong learning culture in which I can not only benefit from others but also contribute and pass on knowledge myself.

Even though my calendar includes a few recurring meetings, no two days are the same. Every morning we have daily stand‑ups where our data team shares progress on projects and tickets, potential issues, or changes in the week’s plan. I usually have two to four meetings a day with Product Management, the managing director, system admins, or colleagues from other products to discuss projects or potential Data Science solutions.

Once a week we have our Data Science meeting (jour fixe), where we discuss current issues in detail, review the status of all our projects, or present and debate new ideas. I also sit down with my team members to discuss projects and prioritize the tickets we’ll tackle in the following week.

I spend the rest of my time designing new projects, helping my teammates with questions and problems, working on tickets myself, or writing new tickets for the coming week. I also test and verify that all running “jobs” (e.g., retraining runs, data exports, automated computations) are functioning as expected.

Overall, I see the biggest changes from AI in the nature of the projects we’re working on right now and in the speed at which we implement them compared to two years ago. A large share of our daily releases now includes new AI‑based features. Enthusiasm for AI’s possibilities is palpable across the company—sometimes it’s even a challenge to convince colleagues that an LLM isn’t always the right solution for their specific problem.

AI is also firmly embedded in day‑to‑day work. I’ve built small personal “agents” that take repetitive tasks off my plate—for example, drafting emails, rephrasing text, finding root causes of errors in code, or sparking new ideas when I’m stuck. At the same time, I don’t rely on AI for important decisions, and I’m very deliberate about which information I share.

Looking at the numbers, it’s very clear to me: we need more women in IT. STEM roles are largely occupied by men, and it’s proven that diverse perspectives and ways of thinking are key to success. That’s indisputable, which is why I believe we need to actively push for change. One way is to introduce these professions to girls already at school. We need mentoring programs and female role models. We have to show that a woman isn’t odd or an outsider if she wants to pursue a career in STEM or any other male‑dominated field. I’ve met so many brilliant girls who were discouraged by prejudice and self‑doubt and ended up studying what was expected of them as women. I think we need to change these behaviors. IT is open to everyone and is a very rewarding field that’s a lot of fun.