Artificial Intelligence Meets Logistics: Data-Driven ‘X-Ray Scan’ for Refrigerated Containers

Dear Joshua, congratulations on your graduation and your outstanding bachelor’s thesis! You tackled a highly relevant topic for the logistics industry, delivering tangible value to your partner company and current employer, Hapag-Lloyd.
What exactly is your thesis about?
I explored whether it’s possible to predict the contents of a refrigerated container using sensor data — essentially a data-driven digital X-ray.

How did you come up with the topic? Was it your initiative or developed together with Hapag-Lloyd?
The idea originated with me. After numerous brainstorming sessions with Hapag-Lloyd’s DIA department, I was determined to choose a topic that was both scientifically exciting and highly relevant. I started by creating a large mind map to capture all ideas and directions. From there, I developed a quantifiable matrix to evaluate criteria such as relevance, feasibility, and scientific merit. In the end, four topics remained, and the clear favorite was the classification of freight types using sensor data.

What is special about it? 
What makes my work unique is that I used real IoT sensor data from cold chain logistics. This data is vast, irregular, and sometimes error-prone. I compared two modeling approaches: a random forest as a classic machine learning method and an LSTM (long short-term memory) for handling time sequences. The results showed that the random forest was more efficient—robust and faster with the given data—although LSTM could offer potential for more complex temporal patterns.

Hapag-Lloyd was among the first shipping companies worldwide to implement comprehensive sensor technology for real-time monitoring of container position, status, and movement. How does your work build on this?
This technology forms the foundation of my work—without live data, such analyses wouldn’t be possible. My thesis demonstrates how this data can be used not only for monitoring but also for intelligent applications such as automatic freight classification. This creates added value: patterns in the sensor data can support operational processes as well as customs clearance and sustainability efforts. In this way, my work reflects Hapag-Lloyd’s vision of turning data into actionable AI-driven applications.

Will containers become even smarter? Could this further optimize supply chains?

My work shows that sensor technology is just the first step. Real-time monitoring is essential, but the real value emerges when you interpret the data and turn it into actionable intelligence. A container can do more than report its location or temperature—it can also reveal what type of goods are likely inside. This shifts the focus from reactive monitoring to proactive decision-making.

How important was it for you not to ‘just’ do research, but to create something concrete and applicable?

Extremely important. I wanted to deliver something that wouldn’t just stay on paper but could provide real value to Hapag-Lloyd. That’s why I chose a topic based on real sensor data with direct practical relevance.

You joined Hapag-Lloyd right after graduation. Did your bachelor’s thesis play a role in securing a permanent position?

After graduating, I continued at Hapag-Lloyd as a software engineer. A key factor in the department’s positive response was my knowledge of machine learning. My bachelor’s thesis was a real door opener: it demonstrated what’s possible with sensor data and gave me strong internal visibility. This now allows me to stay involved in AI and continue developing in my role.

Overall, did your studies prepare you well for your current job?

Absolutely. The project work in Programming 1 & 2 and Project Management taught me how crucial teamwork is. You don’t just learn to code or structure projects - you learn to collaborate with very different people, resolve conflicts, and still achieve shared goals. These are lessons you won’t find in any book, but they’re invaluable in professional life.
Another turning point was choosing Data Science as my minor. That’s where I discovered my passion for machine learning - something that directly led to my thesis and career path. Without that focus, I probably wouldn’t have gone in this direction.
And what really made the difference was Professor Andy Witt. He didn’t just teach knowledge; he encouraged me to carve my own path. That support kept me intrinsically motivated and helped me approach topics with genuine passion.
Looking back, my studies didn’t just teach skills - they gave me clarity about what excites me. That’s what I wish for new students: dare to choose specializations that fascinate you, take project work seriously - it’s a safe space to make mistakes - and find mentors who inspire you. Then your studies become more than a degree; they become a springboard.