Spotlight on Tech with Niklas

Niklas is part of our AI team as an Associate Machine Learning Engineer. His journey at P7S1 began with an internship in our Data & AI department. Since the beginning of this year, he has been a full team member, now working in a dual role as a technical Product Owner and Developer. In his role, he helps our colleagues work even more efficiently. On October 25, he will provide insights into the topic of "GenAI-Enabling" at the ITCS Tech Fair. In our next "Spotlight on Tech," he’ll tell us more about it.

Niklas, "Associate Machine Learning Engineer", that sounds exciting. What can we imagine your role to be?

As an Associate Machine Learning Engineer, I am a software developer focused on artificial intelligence. I train machine learning models on raw data, enabling them to independently recognize patterns and make predictions or decisions. This allows us to support our colleagues in their daily work and optimize the overall program planning of our formats through algorithms.

What motivated you to speak specifically about "GenAI-Enabling" at the ITCS fair?

GenAI has been a hot topic since the release of tools like “ChatGPT", “Midjourney”, “Suno”, and others. This makes it especially important to identify use cases that offer genuine value. In a media company like P7S1, there are numerous interesting applications: from AI-generated social media campaigns to custom-generated content. We’re already exploring these future-focused use cases but also looking at ways to support daily workflows. Through my tech talk, I want to provide a closer look at the development of our GenAI application, discuss the challenges we encounter, and present possible solutions.

Can you give us a quick overview of what your GenAI application does and how it’s integrated into the daily work of P7S1 employees?

Our GenAI application, called the “P7S1 GenAI Playground,” has evolved into a tool that hundreds of employees use on a daily base. With the Playground, our colleagues can not only chat with leading language models like “OpenAI’s GPT-4” or “Anthropic’s Claude 3.5 Sonnet” in a secure environment, but it can also create specialized assistants. These assistants help them with daily tasks, such as generating compelling headlines from TV segment texts for Galileo. Additionally, Playground users can create images via “OpenAI's DALL-E 3” or generate transcripts from audio files using “Whisper.”

What challenges have you faced in developing the GenAI application, and how did you overcome them?

Our top priority is ensuring that our GenAI application, now an integral part of our colleagues' daily workflows, works flawlessly. The challenge is that our application requires a continuous connection with models from providers like “OpenAI.” To address this, we developed automated tests that check model accessibility and alert us in case of any disruptions.

Niklas giving an internal presentation on the "P7S1 GenAI Playground"

Wow, that all sounds really exciting. Let’s go back to the beginning of your career — when did your passion for AI first start?

I studied "Finance & Computer Science" for my bachelor’s and master’s at TU Munich, where I got some early exposure to AI through a few university lectures. Initially, I always thought I'd end up in finance. However, while working on my master’s thesis on “Machine Learning for Corporate Profit Forecasting” with an international investment company, I realized I was far more interested in data and machine learning. That’s when I delved into how I could enter the data and AI field and took the leap from finance to AI. Looking back, I’m very glad I trusted my intuition.

You didn’t initially consider ProSiebenSat.1 as an employer. How did you end up here?

That’s true — I’d never thought of ProSiebenSat.1 as an employer, especially not in Data and AI. I thought, “They just do television.” Then, by chance, I came across an internship posting in Data Science & AI. I applied within five minutes, and two weeks later, the interviews were over, and I had the internship. From day one, I knew I wanted to join the team long-term, so I gave it my all.

That worked out great, as you were hired as an Associate Machine Learning Engineer at the beginning of the year. What did you enjoy most from the start?

The team is very dynamic, highly motivated, and incredibly supportive. My colleagues are genuinely cool, interesting, and diverse— a mix of marathon runners, techno DJs and gamers. The projects we work on are super exciting, and we get to work with a variety of data sources, from classic tabular data to image, audio, and video data. I also appreciate that ProSiebenSat.1 is a modern employer that values work-life balance. And ProSieben has always been the coolest TV channel! (laughs)

Finally, is there anything you’d like to share with anyone interested in a career path in AI?

From my own experience, it’s so important to listen to your heart and pursue your interests. It’s crucial to identify with your work, enjoy it, and see a purpose in it. Also, even in technical roles, being a good communicator is incredibly valuable. You communicate a lot, not only within your development team but especially with other departments, which makes it easier to collaboratively create solutions for their challenges.

Are you interested in exploring our AI department through an internship, just like Niklas?

👏Great news! Here’s your chance...

Intern Data Science & AI (m/f/d)

An