High-Tech Inventor to Further Machine Learning in Science

From the early days of 3D face recognition to great entrepreneurial and academic successes, engineer and computer scientist Alex Bronstein is no stranger to groundbreaking innovations. His path has meandered between academia and industry to find research applications that better the world. The new Institute of Science and Technology Austria (ISTA) professor wants to advance machine learning for life sciences applications. An interview about his unusual career, versatility, and striving for beauty in all its manifestations.

You are a computer scientist, engineer, inventor, and tech entrepreneur. How do you juggle such broad interests?

I view these various interests as synergistic in many aspects. I would describe myself as an engineer with a drive for impactful research and applications that can better the world. To me, industry is very much aligned with academic research, and I’ve used tools from academia to push industrial applications, and vice versa. There is a lot of mutual enrichment, synergies if you like, in the two sectors.

Your research revolves around machine vision and learning. What motivates you to make machines that learn, see, and hear the world?

Endowing machines with perception is a mystery even to the engineers who build them. I see humans endowing machines with agency as a somewhat metaphysical act of creation that allows us to learn about ourselves too. It is a very satisfying feeling.

In addition to founding multiple startups, you were a Principal Engineer at Intel. What marked your spirit during this time?

In the early 2000s, my identical twin brother, Michael Bronstein, and I ventured into making face recognition fully three-dimensional and had to build our own cameras based on structured light. In 2012, we sold our 3D sensor technology to Intel and I joined the company as Principal Engineer to develop it further under the RealSense brand. Then, one day, I saw a TV commercial for the Intel Ultrabook featuring RealSense, the first 3D camera in a portable device. Looking back on how this technology went from a lab prototype to production and how we braved the numerous challenges and brought it to the market, I felt proud of the work we invested for years—even though some of the ideas I’m most proud of remained entirely invisible to the end user. No matter whether and who receives credit for an invention, I find the joy of first discovery one of the strongest driving forces in my life.

How does it work, to invent a new technology?

My team and I are naturally more fit to develop a cool technology first and then ask what it can be useful for. But sometimes, we build a technology targeted to a specific problem and then see how to develop it for broader applications. I have worked with such technologies in biotech, improving in vitro fertilization procedures, developing biomedical devices for various applications, and advising on devices for cardiac surgery. I am very proud of a few aspects of this work, but I also became aware of some dark corners of the medical device market.

You headed the Center for Intelligent Systems at the Technion in Israel while leading your research group. How do you inspire collaborators and team members?

Initial interactions with collaborators, prospective students, and postdocs are very individual, and there is no single recipe for success. In my group, I promote mutual learning, symmetrical interactions, and freedom of exploration. I try to create and maintain a cloud of creativity to help my group members thrive. I remember myself as a graduate student—I often didn’t do what my PhD advisor wanted me to do, but I succeeded in areas that were deemed impossible. Likewise, I constantly encouraged some of my best students to take paths less traveled by and I’m very proud of where those paths led them.

ISTA is not new to you. You have been a Visiting Professor for a year. How well do you feel embedded in the ISTA and xista innovation ecosystem?

A lesson I learned as a PI, founder, and also as a regular investor in tech startups is that nothing is more important than people. At ISTA, I find vibrant minds open to cross-domain collaborations beyond traditional discipline boundaries. I also appreciate how quick and agile the tech transfer ecosystem is at ISTA and xista. It feels like a startup. To me, ISTA is a real example of how science should be done in the 21^st century, a place that fosters cross-pollination and thinking outside the box. This year’s Nobel Prizes in Physics and Chemistry, awarded to machine learning technologies enabling scientific discovery, beautifully illustrate this idea.

What would you like to achieve at ISTA?

I want to push the boundaries of machine learning for applications in the life sciences. I see my role in helping create strategic directions toward structural and cell biology applications, as well as toward single-cell biology. My work will inform ML-friendly experimental design and data collection strategies. The great infrastructure available at ISTA will help us achieve this goal.

Your brother, Michael Bronstein, was recently named the Founding Director of AITHYRA, the new institute for AI and Biomedical Research in Vienna. How do you envision the future of AI research in Austria?

One of Austria’s strengths as a location for AI research is the growing investments in ML research toward various applications rather than merely for advancing ML. With the existing research centers and universities, it will be a lot easier to build these bridges. By investing in bringing in talented people, I can see Austria becoming a center of excellence for AI in science, and perhaps one of the world’s leading locations in five years.

Your interests go beyond science and innovation to encompass music, photography, poetry, writing, painting, gastronomy, and sailing. Is this how you let your buzzing mind wander?

I have always had broad interests which I also sought to cover in my work. I have been passionate about biology, finance, and music for many years. My current research interests cover biology quite well, and one of my recent entrepreneurial adventures in algorithmic trading satisfied my desire to touch finance. On the other hand, AI in music remains a whole new world for me to explore. It intrigues me to tackle how to use live biofeedback of the audience’s perception during a performance. All in all, I would say I appreciate beauty in all of its manifestations, and that goes well beyond science.