With his studies and a trainee program at BMW, Nils Quetschlich made a quick start to his career in the field of autonomous driving. But his desire to do a Ph.D. and experience "pure research" led him to quantum computing, for which he develops software. The young computer scientist is always ready to seize an opportunity and looks forward to new challenges.
By Maria Poxleitner
“When you reach the goal super fast, then it's also a bit boring.” Nils Quetschlich is sitting at his desk in the “TUM Quantum Space”, as his team has dubbed the shared office. The bright, loft-like room is located on the top floor of the Technical University (TUM) building on Arcisstraße, directly opposite the Alte Pinakothek in the heart of Munich. The 30-year-old computer scientist develops software solutions for quantum computing. Working on topics where the solutions are not always immediately obvious and where you sometimes have to grit your teeth – that's something Nils didn’t want to miss in his work. He likes to set himself goals, says the computer scientist. Like his doctorate. “It takes a few years, it's a challenge, and in the beginning you don't know if you're going to make it. But that's why I find it so appealing!”
Taking the perspective of future users of quantum computers plays a key role in Nils' research. “At the moment, it's not that easy to use quantum computers,” says the Ph.D. student, describing the status quo, “you have to be quite an expert.” The aim of his work is to develop methods and tools that enable the efficient use of quantum computers even without such specialist knowledge: “Expert knowledge should not be a requirement.” This has long been the case with today's conventional computers – everyone is using a laptop or smartphone without having to know what is going on at software and hardware level.
The first challenge that users face is choosing the right hardware, explains Nils: “There are many different qubit technologies, and for each technology there are several companies offering quantum computers.” Using the first quantum computer that comes along doesn't necessarily give the best result, says the computer scientist. That decision should therefore be taken away from the user.
Once the quantum computer is selected, the algorithm to be executed must be compiled – just as with a classical computer. “This means that we make the algorithm executable for the selected computer,” explains Nils. The algorithm has to be redesigned in such a way that it satisfies the limitations of the hardware while making the most of its advantages. For example, only quantum gates that are available to the selected quantum computer can be used, as only certain gates are supported depending on the hardware technology.
Position
Ph.D. student
Institute
TUM – Chair for Design Automation
QACI
Degree
Systems Engineering & Automotive Software Engineering
Nils is developing software methods that make quantum computers easier to use for end users. For example, he has developed a program that automatically chooses the most suitable hardware and provides an optimal compiler for the application the user wants to run.
Compiling consists of several individual steps, Nils continues. There are now a number of complete compilation programs, called compilers, which carry out all these steps in one. Companies like IBM or Google provide such programs. “However, the point is that the individual compilation steps are implemented differently by the various programs,” continues the doctoral student, “and since we want the best possible compilation result, we want, that for each step the program is used that performs this step best.” To get the best possible compiler for a specific application, compilation steps from different vendors must be combined. On the one hand, this requires expert knowledge to create the appropriate interfaces between the different programs. On the other hand, all combinations would have to be tested if you wanted to find out for yourself which sequence of compilation steps is optimal. “All in all, this takes a lot of time, is complex and requires a lot of expertise,” summarizes Nils. In his dissertation, he used machine learning to develop a program that automates this process. Depending on the application, the program first determines the appropriate hardware and then provides an optimized compiler for that hardware. With this tool, the user only has to specify the algorithm and ultimately receives a result, explains the computer scientist. “But he or she should not be exposed to all the complexity in between.”
Nils first came into contact with the topic of quantum computing as part of his bachelor's thesis. However, the motivation for this thesis was not the topic itself, but the desire to do his bachelor's thesis abroad. In the end, he went to Hokkaido University in Japan. Nils was given this opportunity by his then supervisor and now Ph.D. advisor Robert Wille, who was still a postdoc at the University of Bremen, where Nils was studying. His bachelor's thesis was already about software for quantum computers, says the computer scientist, adding with a wink: “And then I sort of took a wrong turn.”
After completing his bachelor's degree in Systems Engineering, which Nils describes as a mix of computer science, mechanical engineering, and electrical engineering, he switched to the TU Munich for a master's degree in Automotive Software Engineering. During his undergraduate studies, it quickly became clear that computer science was his favorite subject, while at the same time he had always been fascinated by the technology that goes into cars. His two wishes, to switch to a different university – “as a challenge, that was important to me” – and to move to a new city, were also fulfilled when he decided on this program. Nils wrote his master's thesis at BMW in the field of autonomous driving and subsequently, went through a multi-stage selection process, to secure a place in a BMW trainee program, also in the field of autonomous driving. “The program was a perfect match for my studies.” With this acceptance, the decision was made for Nils, who had initially wavered between a Ph.D. and a direct entry into industry. The 18-month program was a lot of fun, says the computer scientist, and he could also spent two periods abroad. The trainee program was accompanied by a permanent employment contract, and so, after the 18 months, Nils directly started as a technical product manager in a management role.
For one and a half years – because: “The thought of doing a Ph.D. never went away.” In order to find his “peace of mind,” as Nils says with a laugh, the computer scientist began to look around for possible doctoral positions. An industrial doctorate was out of the question for Nils – if he was going to do a doctorate, he wanted to do “pure science”. His former bachelor's thesis advisor, Robert Wille, promptly offered him a position. Nils says he didn't even think about doing a Ph.D. in quantum computing at first. “I thought it was all so theoretical and didn't fit in with my automotive stuff.” But the fact that Robert Wille, who had made him feel so well taken care of at the time, would move to Munich a few months later for a new professorship, that is to the city where Nils wanted to stay, ultimately provided the young computer scientist with a coherent overall package: “I didn't have too much prior knowledge, but I had a very positive gut feeling.” His fundamental interest and open-mindedness did the rest: “I can get excited about a lot of things!”
The topic of his doctoral thesis was not as theoretical as he had initially feared. The “Munich Quantum Toolkit (MQT) Predictor”, as Nils and his team call the program he developed, is very interesting for the Leibniz Supercomputing Centre (LRZ), for example, Nils points out. As a computing center, one of the basic tasks of the LRZ is to provide computing resources to its customers. As part of Munich Quantum Valley, the LRZ's quantum department wants to expand quantum computing as a general service and already has several quantum computers based on different hardware technologies on site. “Users send their calculations to the LRZ, but they probably don't care which quantum computer they are running on,” explains Nils. Therefore, the question addressed by the MQT Predictor is particularly relevant for the LRZ. As part of Munich Quantum Valley, he is in contact with scientists at the LRZ. Getting immediate feedback from scientists who actually want to use and test his program has been very helpful for his work, says the doctoral student. When he's not in meetings with other research groups, he spends most of his time in front of his computer screen. To compensate, Nils likes to play the piano. He started during his Ph.D. – “I wanted to learn an instrument.”
Nils is now in the final spurt. He wants to submit his thesis soon. Looking back on the past few years, Nils is very glad that he didn't do his Ph.D. right away, but went to industry first: “I think I would have seen it more as a means to an end, tried to finish as quickly as possible and enjoyed the time less.” At university, you have a lot of freedom that you don't have in industry, he says. “I'm aware of the advantages because I've already had the experience of working in industry.”
After his doctorate, however, he would definitely like to work in industry again. He would like to stay in the field of quantum computing, but it is not absolutely necessary – as he says, he is enthusiastic about many things. Also the “where”, whether he will stay in Munich or whether he will be drawn out into the big wide world again, is still open. What would speak for Munich would be the proximity to the mountains, because what Nils enjoys in his freetime is paragliding. When time and weather permit, he straps his equipment onto his back and heads up the mountain. He doesn't see it as an action-packed hobby, but as a way to relax: “In the air, I'm by myself.” He particularly likes the Wallberg. Start running, speed up, take off – and then he enjoys the beautiful view of Lake Tegernsee.
Published 28 March 2025; Interview 6 February 2025