Klaus vom Hofe


Reboot with algorithms: Markus retrains

Everything back to the beginning: Markus van Ballegooy has turned his job at Deutsche Telekom upside down. Innovations such as artificial intelligence are doing the same to his area of expertise. So he jumped in at the deep end and retrained. A portrait.

Markus van Ballegooy during project work.

Flashback to summer 2020: good mood during project work. On this day, the task was to solve a team task in a group of four and share it with others via video conference. "There was no shortage of fun while learning," says Markus van Ballegooy.

"Users" - the unknown beings. How do they deal with digital applications? In the car, for example, when they see multifunction displays instead of speedometers. Or when they have to deal with voice control for the first time? Companies that are familiar with this will win customers. Because they design applications that can be operated intuitively, solve problems or even provide new, suitable offers. In other words, many things that simplify everyday life.

"The ingredients for this recipe for success are all there today," says Markus van Ballegooy. "There are data volumes, computing power and artificial intelligence. Together, these make it possible to predict, for example, what website visitors intend to do. Or voice controls become possible that respond to naturally spoken words instead of just individual keywords."

The Deutsche Telekom employee immersed himself in artificial intelligence (AI) during a six-month internal training course. He learned the Python programming language and how to use it to develop and apply algorithms. For example, to sift through 30,000 data records with movie reviews from viewers at the push of a button. That is, star ratings as well as texts. The exercise was to find out whether star ratings of movies could also be predicted from written movie reviews. "That's exciting, but it's also time-consuming. For example, when it comes to preparing data - what's called pre-processing. Because the raw text data is mostly unstructured, full of errors and filler words, you have to teach the software those first."

Another thick board of advanced training: artificial neural networks, the supreme discipline of artificial intelligence. "We crammed to get through the tricky topics," says the 49-year-old, summing up the months that lay behind the 20 participants in the "AI / Data Science" course. The moment when they all held their certificates in their hands was all the more beautiful. Shortly thereafter, Markus van Ballegooy joined the Magenta Voice team. This team is working to ensure that as many Deutsche Telekom products and services as possible can be operated by voice in the future. He is helping to develop new neural networks. The aim is to make it easier for these networks to learn and provide more reliable answers in voice dialogs. "For us newcomers, it's now a case of learning by doing," he says.

The eternal "human-technology issue"

Flashback: Markus van Ballegooy was already working on projects for industry, such as speech recognition and multifunction displays, during his psychology studies 25 years ago. Since then, his hobbyhorse has been methods and statistics for more "user experience" - or as it was called back then: "cognitive ergonomics". He prefers to call it the "human-technology issue," i.e., positive experiences when dealing with digital products. The interesting thing from his point of view is that it remains topical. For example, voice dialogs with machines: "Even if these are getting better and more familiar: Many users still expect too much from it or don't get along with it right away." 

In 2001, van Ballegooy moved from university to the Deutsche Telekom subsidiary T-Systems. Back then, he was already working on voice portals for customers of what was then "T-Mobile Deutschland," and later he optimized systems there that hotline agents work with. At the Telekom Innovation Laboratories (T-Labs), he developed methods that companies use to involve users in idea and product development, such as workshop formats with customers.  

Artificial intelligence: jumping in at the deep end

"But at some point I realized: the time was ripe to change," the 49-year-old looks back. A sabbatical year was a first expression of this. Privately, he took a new approach and began training as an osteopath. And he became increasingly fascinated by the AI hype. Even in his studies, he was interested in dealing with large amounts of data. "Then at some point I realized: it's still about the same thing, only that AI enables much more on a large scale - it's like a logical progression for me," says van Ballegooy. "We used to calculate how likely some would buy a product based on 1,000 website visitors*. With AI, we can predict much more accurately what they want at the push of a button based on 100,000 records. It's similar for speech recognition." 

A common thread, then. Markus van Ballegooy advises anyone who is flirting with such advanced training that they already have practical knowledge. "It's a jump into the deep end in any case." Even though it was years ago, he benefited from programming exercises at university. Helpfully, the course participants continue to support each other to this day, after their fresh start in various specialist teams. Markus van Ballegooy is also impressed by the culture of sharing in the new team as well as in the AI and software community of the Deutsche Telekom Group - and by the appreciative way people treat each other. "Whether proven experts or career changers: Everyone is aware that the mix of our different perspectives, experiences and backgrounds leads to really good results for customers."  

Markus van Ballegooy's new path is also an expression of the transformation of the entire IT and telecommunications industry. New technologies are enabling new digital applications for industry and consumers alike. Ideas and developments are creating new momentum in the markets at a rapid pace. Global corporations, so-called hyperscalers, are rapidly making these big - as digital applications for millions. At the same time, Deutsche Telekom is restructuring its networks. The challenge for these networks is to meet future requirements intelligently and flexibly and to quickly incorporate new services, including those from partners. For all of this to succeed, expertise in software, cloud and artificial intelligence (AI) is needed. Equally important: a culture of curiosity and trial and error, openness, cooperation in partnership and agile working. 
Deutsche Telekom is taking its employees along with it on this journey of change. The company has established relevant academies for colleagues in its IT, technology and innovation units, in addition to other training programs. Since 2018, colleagues have been training to become Scrum Masters, for example, as well as specialists in software, AI and data science, under the heading "New Skilling". They do this full-time over several months before contributing their experience to specialist teams and projects and continuing their practical learning. More than 800 Deutsche Telekom employees have now completed these academies. More will graduate this year.  



With our products and our network, you can be sure to always #takepart in all the latest developments and technologies.