AI Research Scientist with a passion for pushing the boundaries of machine intelligence.

Hi, I’m Frederik Berenz,

My journey began with a strong fascination for mathematics and physics, which led me to study mathematics after high school. However, the opportunity for financial and geographical freedom brought me to Mercedes-Benz, where I pursued a dual degree in mechanical engineering in cooperation with the company.

In 2016, I joined Mercedes-Benz full-time as a drivetrain designer for all current gasoline engines in the field. Over time, I took on greater responsibilities as an engine strategist and module team lead for all drivetrains.

Eventually, I felt the need for a new intellectual challenge. During the pandemic, an old MATLAB subscription introduced me to AI, and I quickly became captivated by its capabilities—and limitations. I started questioning:

What can’t AI do? Where are the true limits(assuming high-quality data is available)?

Motivated by these questions, I enrolled in a part-time MSc program in Artificial Intelligence at the University of Bath in 2021, completed it in 2023, and officially graduated in 2024—all while transitioning within Mercedes-Benz to a new role in AI research.

Currently, I work as an AI Research Scientist, specializing in Foundation Model Multi-Agent Systems aimed at creating deeply personalized customer experiences. A core focus of my work is exploring and understanding the limits of Foundation Model Multi-Agent Systems—not just what they can do, but what they inherently can’t.

While foundation models may convincingly imitate AGI in some domains, they’re fundamentally not AGI. Token prediction alone isn’t the final answer. That’s why, alongside developing state-of-the-art multi-agent systems, I’m deeply interested in exploring alternative architectures and learning paradigms to enable more holistic, capable machine intelligence—beyond the current frontiers.

My Linked In Profile URL is: www.linkedin.com/in/frederik-ber. It didn't let me save it