AIT India 2025/26: Nuorail
Meet Juxhino Kavaja, one of the ten participants in the 2025/26 edition of the Academia-Industry Training India programme.
Interview with Juxhino Kavaja
“nuorail develops energy-optimal autopilot software harnessing the power of applied mathematics. Our software minimizes the energy usage of train operation by optimizing traction and regenerative braking forces in real time, while ensuring on-time arrival and compliance with operational constraints.”
What problem does your startup solve, and why is it important to you personally?
Controlling a train’s traction and braking forces may sound simple – but doing it in the most energy-efficient way is a very complex problem. Track slopes, curves, tunnels, and even weather affect the optimal driving style. Currently, most trains across the world are manually operated, with driver advisory systems offering suggested speed targets via a display. However, this traditional approach has a significant impact on the energy consumption of railway networks. Indeed, human drivers are unable to execute consistent driving profiles, and most algorithms behind driver advisory systems introduce unnecessary oversimplifications to the problem. As an attempt to overcome these limitations, the industry is steadily, if slowly, transitioning to automatic train operation. However, at present, the algorithms in use are mostly based on heuristic (or “good enough”) methods.
nuorail develops energy-optimal autopilot software harnessing the power of applied mathematics. Our software minimizes the energy usage of train operation by optimizing traction and regenerative braking forces in real time while ensuring on-time arrival and compliance with operational constraints. Our solution combines the experience of a human driver with the consistency of a computer – and goes a step further. Our unique software solves extremely complex planning problems in fractions of a second, delivering energy savings while keeping trains precisely on schedule, even under challenging conditions.
Improving the energy efficiency of rail transportation is important to me because of its environmental and social implications. Higher energy efficiency leads to less greenhouse gas emissions and improved sustainability. Furthermore, it allows railway operators to reduce transportation costs, making train transportation more accessible and attractive for passengers, a topic which is very important to me.
What inspired you to become a sciencepreneur, and what has been your biggest “aha!” moment so far?
I like to believe it was a natural process driven by curiosity, successful results and the desire to have a positive impact on the real world.
After earning a PhD in Control Theory, I was looking for the right field to apply the knowledge acquired during my studies. While the potential applications were endless, I was struggling to find one that motivated me. It is only when I moved to Switzerland that I realized that railways are not the 200-year-old stagnant industry that I had in mind and, after personally experiencing their advantages on a daily basis, I decided I wanted to play my part in making them more accessible and attractive.
In 2023, I joined Dimitris and Ishan at FHNW. By that time, they had already made remarkable theoretical progress in designing advanced techniques for controlling trains in an energy optimal way. With the support of NCCR Automation, and in collaboration with STADLER and Schweizerische Südostbahn AG (SOB), we started a pilot project aimed at verifying the achievable energy savings of our algorithms through real-world experiments under normal passenger traffic.
Every engineer has probably experienced a gap between theory and practice: sometimes the best solution on paper might not be the best one in the real world. Theoretically, the control algorithms we had developed presented the best approach to address the energy efficient train control problem and while simulation results confirmed this, we were not sure if in practice this would have worked out.
My biggest “aha” moment so far took place during one of our first test drives where we confirmed that our algorithms achieved a remarkable improvement in energy efficiency. At that moment, it became obvious that we had built a solution capable of defining new standards of energy efficiency for train operations, allowing for energy savings of more than 20%.
What unique perspective does your academic background bring to your startup?
At nuorail, we believe that control theory is crucial for addressing several challenges faced by railway transportation. This is why all our founders have a background in this discipline. Additionally, our energy-optimal autopilot software is built upon fast real-time control libraries which were developed by Dimitris, our CEO, during his PhD studies.
A famous paper in our research community suggested that control theory is a “hidden technology”: while control techniques and algorithms are pervasive, their existence is typically ignored by the end user – and sometimes even by engineers of related fields.
Driving a train in an energy-optimal way is an example of a special class of problems known as optimal control problems. These types of problems have been studied by the control community since at least the 1950s. Despite this, many train manufacturers fail to recognize this and miss out on the many existing theories and algorithms developed by our community.
While most train manufacturers address the energy efficient train control problem with ad hoc and heuristics methods, control theory expertise has helped our startup to approach the same problem in a mathematically and physically rigorous way, thereby leading to a much more complete and effective solution.
What’s one surprising lesson you’ve learned since launching your startup?
Recognizing when something is good enough. In academia, we typically pay lots of attention to details and we often aim for perfection. While this is certainly important at certain stages, it can become a burden in others. Our startup has taught me to recognize when things are good enough. Rather than always seeking perfection at all costs, I learned to recognize when sufficiently good results are enough, especially when they can enable constant progress.
If you could host a dinner with three innovators (past or present), who would they be and why?
Alberto Broggi, the General Manager at VisLab (now part of Ambarella Inc.) and former professor of Computer Engineering at the University of Parma, Italy. He pioneered the use of computer vision for automotive applications and particularly driverless cars. I’d be eager to discuss over dinner about if and how railway automation can learn from road automation.
Giovanni Marro, recognized as one of the most influential researchers in the field of automatic control. He was a professor of Control Engineering at the University of Bologna, Italy, and developed the so-called “geometric approach to control theory”, which opened the door to a deeper understanding of our discipline and to many new control results. I would like to hear his honest opinion regarding newer developments that our research community is chasing, such as learning-based and data-driven control techniques.
Alberto Sangiovanni Vincentelli, a full Professor of Electrical Engineering and Computer Science at the University of California, Berkley, and co-founder of Cadence Design Systems and Synopsys. He has shaped the digital landscape of today through research and technological innovations. I would like to invite him to learn about what he thinks is the secret to a successful entrepreneurial venture.