From Mathematical Uncertainty to Smart City Innovation with Data Analysis, Machine Learning, and AI. An Interview with Stan Sobolevsky

Prof. Stan Sobolevsky

Researcher, professor, and entrepreneur. ML/AI researcher/engineer at Meta and Research Professor at New York University.

Mathematician by training utilizing the value of big data, networks, machine learning and AI for urban and business analytics and innovation.

Passionate in applying analytic mindset and problem-solving skills to exciting scientific challenges and cutting-edge industrial problems with those who think beyond the code.

Stan Sobolevsky at the DIgital City Lab, Institute of Computer Science, Masaryk University. Photo: Irina Matusevich

Can you tell us about how your professional journey in digital urban innovation began and how it led you to founding your own lab?

My journey into this field began around 15 years ago at the MIT Senseable City Lab, where I spent several formative years. We were among the early pioneers in the digital transformation of smart cities—at a time when large-scale urban data was just becoming accessible. After that, I moved to New York University, where I founded and led the Urban Complexity Lab. Over time, the lab gained recognition—publishing in leading journals and collaborating with industry, public institutions, and urban agencies.

Explain to us how the Urban Complexity Lab moved to become the Digital City Lab, and what it is bringing to Brno, Masaryk University and its Institute of Computer Science?

For about five to six years, the Urban Complex Lab evolved and grew through this work. Then, in 2022, with the generous support of Masaryk University—particularly through the MUNI Award in Science and Humanities (MASH)—we were given the opportunity to establish a new lab in Brno.

This new chapter marks the transition from the Urban Complexity Lab to what is now the Digital City Lab. Our goal is to bring this research to the center of Europe and open it to even more talent—both from across Europe and globally.

Has the Digital City Lab at Masaryk University been successful in attracting talented PhD students?

Yes, it has been quite a success. In just a few years, we’ve attracted a strong cohort of PhD students—one has already completed and defended their thesis and is now continuing with us as a postdoctoral researcher. At the moment, we have five PhD students actively engaged and another two candidates awaiting formal enrollment, which makes the lab a substantial and vibrant research environment.

Most of our core activities are now based here in Brno, even more so than in New York. This shift has been supported by our growing collaborations with local industry partners who are deeply engaged in our work.

Photo: Irina Matusevich

How are you collaborating with industry partners, and what are your goals for these partnerships moving forward?

We’re now working closely with a range of companies, and we’re excited about evolving these relationships into a full-scale industry–academia consortium. Our mission is to address real-world challenges of digital cities—leveraging urban data for innovation and co-developing practical, data-driven solutions with our partners.

But that’s not all—because beyond research, the Digital City Lab is also deeply engaged in education, through its Digital Talen Lab activities, as Datathon. If someone values experiential, informal learning—and believes that gaining skills in data analytics, AI, and machine learning is essential in today’s world—how does the Lab provide access to that kind of education?

We see the Digital Talent Lab as a natural extension of our work at the Digital City Lab. It’s designed as a space to attract and develop new talent—people who are curious about digital urban innovation and want to understand how it functions in real-world settings. Participants gain hands-on exposure to the key skill sets needed today: data analytics, machine learning, and AI, all within a professional context.

Photo: Irina Matusevich

Specifically, how do the Datathon offer free, hands-on learning opportunities? And who are these programs designed for? Can students, professionals, or even complete beginners take part in order to stay current with fast-evolving digital technologies?

Datathon isn’t a traditional academic program. It’s an experiential learning environment. While passing entry trial test or show their proficiency by their resume, participants are actively involved in projects—with us or with our industry partners. They build practical skills, gain real experience, and start growing their professional networks. Importantly, they’re mentored by both academic researchers and industry advisors. That dual perspective offers real value—not just in terms of learning, but also in helping shape meaningful career paths. So yes, the Digital Talent Lab offers a deeply immersive and valuable experience for anyone aiming to grow into the digital future of urban innovation.

Let’s take a step back to your personal journey. As someone with a background in mathematics, how did you find your way into the world of data analytics, machine learning, and AI?

Well, my background is in mathematics—that’s where my formal training began. For the first decade of my professional academic career, I was working in a very narrow area of differential equations. I was super excited about it, as were maybe another 10 people in the world, but the rest of the world pretty much didn’t care—which often happens in theoretical science. I enjoyed the intellectual challenge for some time.

Then I realized that if I wanted to create real impact, I probably needed to try myself in more applied domains. Back then, data was undergoing its Golden Age—vast amounts of data started becoming available, with very few constraints on their use. Many industrial companies, banks, telecoms, insurance companies, and transportation firms were very open to sharing data with academic partners.

Photo: Irina Matusevich

You mentioned a pivotal moment in your career—your time at the MIT SENSEable City Lab. How did that experience shape your path, and what led you to eventually start your own lab?

I got to know Professor Carlo Ratti, who was leading the MIT SENSEable City Lab. After a couple of visits to get to know the lab and the team, I decided to stay for good—for the next three years, as it turned out. Then I started my own lab. That was an interesting transition in the middle of my academic career, which led me to where I am now in both academic and professional worlds.

How are you enjoying this transition from being a mathematician, and what is fascinating for you about data or related fields?
As a mathematician, my research curiosity has always been inspired by uncertainty and indeterminism. Traditionally, mathematics deals with purely deterministic systems—mathematical laws are usually 100% robust, and the systems they describe tend to be deterministic. For example, if you know the exact state of a system at a given moment, you can predict its future state with great accuracy. Take celestial bodies: by observing their current positions and knowing the laws governing their motion, you can precisely calculate their locations centuries into the future.

Can you explain us the mathematical concept of uncertainty?

That’s often how the micro-world operates, but not the world we live in as social beings. We are surrounded by uncertainty—not just perceived uncertainty but fundamental uncertainty. There are two types of uncertainty: one is perceived randomness, which arises because we lack complete information or computational power. For example, when you toss a coin, it seems random because it’s hard to predict its trajectory perfectly. But if you had perfect knowledge of the initial conditions—speed, angle, air flow—you could calculate exactly how it will fall.

The other type is fundamental randomness, like in quantum phenomena, where uncertainty is inherent. For example, you cannot simultaneously know both the exact position and momentum of an electron. This fundamental indeterminism is actually how our brains work, since neurons operate in the quantum world. This means our thinking itself is influenced by uncertainty at a fundamental level.

What sparked your interest in indeterministic systems, and how did that lead you to study complex human behaviors like mobility and crisis response?

From the beginning of my career, I worked on systems that were not traditional mathematical objects. For instance, my PhD thesis was devoted to differential equations but of a special kind, which unlike usual cases having a unique solution given initial conditions, the system could have infinitely many solutions branching out. This models a more realistic indeterministic world where physical laws govern evolution but at many points there are choices or multiple possible pathways—much like how we make choices in life within certain constraints but with freedom and responsibility.

That fascination with uncertainty and indeterminism has driven me to move into increasingly uncertain areas of science, such as modeling human mobility. These systems are influenced by physical constraints like air flows but also contain elements of chaos and unpredictability. This blend of order and uncertainty is what makes the field so exciting for me.

What led you to focus your research on cities as complex, human-driven systems—and what makes them such a fascinating subject for scientific analysis?

Over the past 15 years, I’ve been focused on what might be the most indeterministic system we know—human society, and in particular, our cities. Cities are incredibly complex systems, shaped by the actions of millions of individuals, each behaving in unpredictable ways.

And yet, when these interactions come together, they form patterns. While the actions of a single person may be random, the system as a whole isn’t. Cities follow certain laws and regularities that can be observed, analyzed, and—perhaps—better understood. Uncovering those patterns is both fascinating and intellectually rewarding.

Is it ever daunting to work with something so unpredictable—or do you enjoy that challenge?

Every powerful tool comes with its risks. That’s where responsibility becomes critical. For instance, when we first began working with large-scale datasets—like national-level records of phone activity or credit card usage—even anonymized data carried risks. With enough effort, someone could potentially infer personal details. That’s a serious concern. Today, access to such data is much more restricted, and there’s a stronger focus on privacy. The more we’ve learned about working with this kind of data, the more capable we’ve become at safeguarding it.

There are now strict regulations in place, and ongoing research is helping us refine the boundary between useful insights and ethical data use. It’s a delicate balance—but one that’s essential as we continue to explore the future of smart cities.

What types of data are you able to extract for your research, and what are the legal and ethical boundaries you must observe? How do you ensure compliance with data privacy regulations while leveraging this data?

Knowing the regulations, following the guidelines, and ensuring technical security for the data—that’s the necessary price to pay to be able to leverage data for innovating our cities. Making cities smarter, more informed, and more efficient, and operating them at the scale, care, and pace required in today’s world wouldn’t be possible without interdependent solutions like machine learning and AI.

We face paramount challenges such as sustainability, congestion, and efficiency that need to be solved. Fortunately, the solution comes with the vast amounts of data we produce and the digital technologies now available—machine learning and AI—that help us extract valuable insights from data that simply wouldn’t have been accessible 10 or 20 years ago.

We have these powerful tools, and there was a real need to use them. But, like with every powerful tool—nuclear power, for example—responsible control and caution need to be exercised. Balancing innovation with privacy and security is critical, and that’s something we take very seriously.

For those inspired by your journey and eager to develop their own skills in data, AI, and Machine Learning, what would you say to encourage them to join the Digital Talent Lab community?

I would say—don’t hesitate! The world of data, AI, and Machine Learning is moving incredibly fast, and there’s never been a better time to get involved. At the Digital Talent Lab, we welcome curious minds—whether you’re a student, a professional, or simply passionate about digital innovation. Here, you’ll have the chance to work on real projects, learn from both academic and industry experts, and become part of a vibrant, supportive community. If you’re ready to challenge yourself, grow your skills, and help shape the future of digital cities, we’d love to have you with us.

Thank you for the interview, I will be happy to come with more questions.

Zuzana Jayasundera

Photo: Irina Matusevich