Cashless transactions are making deeper inroads into our society by seamlessly authenticating different financial systems in a secure and fair manner, thus providing equitable access to all, including the underserved sections.

Sunpreet Singh Arora, our next pathbreaker, Lead Research Scientist at Visa Research (California), leads a team that creates cutting-edge identity and authentication capabilities using biometrics, applied cryptography, machine learning and other state-of-the-art technologies.

Sunpreet talks to Shyam Krishnamurthy from The Interview Portal about his work on design and fabrication of 3D fingerprint models which helps better understand fingerprints as a biometric trait in order to benchmark different fingerprint sensors and systems.

For students, technologies of today have made our daily life much easier, but these innovations come from highly specialized roles where in-depth research and development expertise in a specific domain is required !

Sunpreet, can you tell us about your growing up years?

I grew up in a typical middle-class household in Gurgaon, Haryana. My father worked as a mechanical engineer and subsequently as a manager at a major automobile manufacturing company, while my mother was a homemaker focused on running the household and raising me and my brother. While in middle and high school, I was mostly inclined towards academia. I loved to read books on various topics from science to fiction, had an insatiable curiosity to learn about different subjects (from physical sciences and mathematics to social sciences), and participated in quizzes. I was also interested in cricket and music as passionate hobbies. 

I distinctly remember when my dad bought the first desktop computer. I was in 8th standard at that time. My first exposure to a computer led me to learn how to design websites, and subsequently computer programming. It piqued my interest to a great extent in pursuing computer science.

What did you do for graduation/post graduation?

I hold a B.Tech. (Hons.) in Computer Science and Engineering from IIIT-Delhi and a PhD in Computer Science from Michigan State University (direct PhD without Masters).

What were some of the influences that led you to such an offbeat, unconventional and unique career?

Given my deep interest in design and development of software in high school, I really wanted to pursue computer science engineering for my undergraduate studies. So, I gladly took the opportunity to learn the fundamentals of computer science at IIIT-Delhi. 

IIIT-Delhi has a strong academic research culture given all faculty have their PhDs from top institutes across the globe. The culture, combined with my fascination with image processing, computer vision and pattern recognition, and mentorship from Prof. Mayank Vatsa and Prof. Richa Singh, led me to pursue biometrics research for my undergraduate thesis. I worked on addressing covariates that impact performance of iris recognition systems in cases that involve use of different cameras, cataract surgery, and alcohol consumption. In case the readers are not familiar with iris recognition systems, India’s Aadhaar leverages iris biometrics (alongside fingerprints) for uniquely identifying an individual. Iris recognition is one of the foremost identity recognition techniques using a scan of the iris of an individual’s eye. Some smartphones (e.g., Samsung Galaxy) also use iris recognition for user authentication.

After the first two years of my undergraduate studies, I was pretty set on pursuing software engineering as a profession. However, once I experienced industry internships in software engineering, I realized I would not be happy as a software engineer in the industry. I figured that most engineering roles, especially early career ones, do not give one the freedom to think independently and embark on novel pursuits. 

Working on my undergraduate thesis with Prof. Mayank Vatsa and Prof. Richa Singh in image processing, computer vision and biometrics, led me to develop a strong liking for academic research culture, especially the freedom to think and pursue novel research projects. Subsequently, I interned in Accenture Technology labs and liked the work culture and environment of an industrial research lab.

Hence, I decided to seek research roles in industry, and landed a research engineering role at IBM Research India. However, fate had something else in store for me. Close to graduation, my undergraduate advisors made me aware of the opportunity to work with Prof. Anil Jain, at Michigan State University for my PhD. I figured it was a once in a lifetime opportunity to work with and learn from the best – so I convinced my parents to let me pursue graduate school in the US instead of joining IBM Research.

Tell us about your career path after your bachelors

During my PhD at Michigan State University, I had the opportunity to work with one of the top computer scientists, Prof. Anil Jain, in the field of biometrics. Under his mentorship, I worked on tackling challenging problems in fingerprint recognition.

In the first year, I designed and developed a better performing latent print matcher. Latent prints are left when an individual touches or holds an object. Law enforcement agencies collect them at crime scenes and use them to identify suspects by comparing them to legacy fingerprint databases.

Subsequently, in collaboration with my colleagues and our hardware partners, I also developed an end-to-end system for child fingerprint recognition which could successfully identify even a 6-month-old child! Such a system could potentially reduce child swaps at hospitals, help track child immunization records digitally as well as help identify lost or missing children. 

The third major contribution of my work was to design and fabricate 3D fingerprint models. This fundamentally helps better understand fingerprints as a biometric trait and benchmarking different fingerprint sensors and systems. This also helped us study vulnerabilities of fingerprint systems to artificially generated artifacts.

In the third year of my PhD, I also had the opportunity to intern at an iris recognition startup, DeltaID where I developed methods for finding eye region in unconstrained imagery captured using different iris cameras, detecting and correcting gaze in off-angle iris images for unconstrained iris recognition, as well as identifying the physical phenomena that changes the iris structure when pupil dilates/constricts under different lighting conditions, and consequent impact of this change on the performance of iris recognition systems.

How did you get your first break?

I’ve typically followed my inner voice and passion for deciding what next in my career. 

While my PhD experience was fantastic, I realized that most academic research never sees the light of day in real world systems. I wanted to apply my research skills to solve real world problems. Hence, I decided to look for industry research roles after my PhD. 

My mentors in different stages primarily helped me land my next break. During undergraduate studies, my advisors made me aware of an academic research career and helped me land the opportunity to pursue my PhD. During my PhD, my advisor connected me with my first hiring manager who was deeply interested in my expertise in biometrics. So, he flew me in for an interview with Visa in San Francisco, and I landed my first role as a research scientist in the industry!

Subsequently, for different roles within Visa, my colleagues and the “Visa” culture to help each other succeed in their career has helped tremendously in obtaining subsequent opportunities. 

Over the past 6 years, I have been in a few different research roles of progressive responsibility at Visa. I currently lead a research team in Visa Research that is researching and developing next-generation identity and authentication capabilities for Visa, biometrics being one of them. I also lead Visa’s participation in different standard organizations in the identity and authentication space such as Fast Identity Online (FIDO) alliance, and OpenID forum.

What were some of the challenges you faced? How did you address them?

Challenge 1: How to know what to pursue next?

Like anyone else, I did not know that I wanted to pursue a research career till the 3rd year of my undergraduate study. To determine what next, I connected with different professors within IIIT-Delhi and determined which domain excited me most. Also, experiencing software development in the industry via internships helped me eliminate the software engineering path. I was also wavering a bit between a technical degree and a business degree in graduate school. My curiosity and desire to learn the subject material in depth made me choose the former. 

Challenge 2: How to plan and execute to get to the next step?

I believe in keeping things simple. So, I decided to solicit mentorship from professionals in my career of choice – in this case professors who were considered experts in the domain I liked. I started following their advice on reading research papers and working with them to conduct research and publish in reputed venues. This way I was able to network with more researchers and caught the eye of my eventual PhD advisor who offered me a research assistant position.

  • Challenge 3: How to balance career and personal life? 

A research career can be very rewarding yet consuming. During my PhD, I worked extended hours (12-14 hours a day) for months at a stretch without any break. While it increased my productivity initially, I soon realized I was burning myself out, and becoming unidimensional. Hence, I decided to carve specific time out to meet friends and acquaintances regularly, and work on my other passions and hobbies. This helped me rejuvenate and come back with renewed energy to tackle difficult research problems as well as become a well-rounded individual by learning about ongoing work in other domains.

Where do you work now? Tell us about your current role

I currently work as a Lead Research Scientist at Visa Research where I lead a team that creates cutting-edge identity and authentication capabilities using biometrics, applied cryptography, machine learning and other state-of-the-art technologies. Visa Research is a research and development organization within the broader technology organization at Visa.

Research scientists in my team have in-depth research expertise in authentication, identity privacy and security as well as machine learning. Example problems we solve: ‘how does an entity like Visa ascertain if a bad actor is performing a payment transaction?’, ‘how to authenticate a Visa cardholder before or during a transaction in a secure and seamless manner?’ 

What are the skills required for your role? How did you acquire them?

Research scientist roles, in general, in both academia and industry are highly specialized roles where in-depth research and development expertise in a domain is required. The easiest way to acquire such expertise is by pursuing a research-oriented degree such as a PhD, or equivalent research experience in academia or industry. In the industry, research scientists typically craft research agenda in their area of expertise by identifying hard problems to solve, for advancing science and developing methods to tackle those problems. Subsequently or in parallel, they work with business partners such as product managers to find avenues for applying their research to either develop new products or enhance existing products for the company.

One of the things that I love about my job is that my day-to-day is typically very dynamic.  Activities range from meeting my team, working with other internal and external collaborators, reading new research papers, brainstorming about how we should solve research problems, to determining what we should be working on as a team next.

How does your work benefit society?

My work directly impacts financial inclusion. By issuing digital identities to the underserved, we can get them access to financial services for their economic advancement. Furthermore, by making it very seamless to authenticate different financial systems with biometrics in a secure and fair manner, people can equitably access financial systems in future.

Tell us examples of a some specific memorable works you did that are very close to you!

During my PhD, two law enforcement agencies – Lansing police and Michigan State University police reached out to our research lab to seek help on a homicide case where the victim’s phone was locked using his fingerprints. They had read about my work on creating 3D replicas of fingerprints from 2D images. I was able to apply the research I was conducting for my PhD thesis to recreate the victim’s fingerprints and unlock the mobile device to help them gather necessary evidence on the case. As a PhD student, seeing the real world impact of my research was very memorable. Ultimately, I was awarded civilian citation and honorable mention respectively by the two police departments.

Another memorable work is the project on handsfree payments at Visa where we built protocols and systems from scratch for enabling a consumer to checkout directly using biometrics (e.g. using face and/or voice) in different environments without interacting with a credit card or a mobile device. This work had strong internal impact leading to multiple internal and client pilots, as well as external impact – one of our research papers received best demo award at a premier biometrics venue, International Joint Conference on Biometrics (IJCB) 2020.

Your advice to students based on your experience?

The hardest problem is deciding what one wants next in his/her career. It is important to both self-educate and connect with a myriad of people to determine what piques your interest and what is a viable career path for you. The framework of Ikigai helps a lot when making such decisions.

I would also highly recommend actively seeking mentorship from professionals in one’s career of choice, as well as building relationships with colleagues. In my opinion, it goes a long way in opening new career pathways/opportunities.

Future Plans?

I want to become an entrepreneur next, in order to make a bigger dent in the world and ultimately have even more autonomy and control over my time. I am still working on my plan on how best to go about it.