From an offshore engineer to an AI scientist in medicine: life’s like that

Neel Das, PhD

Data scientist | AI/ML in pharma and healthcare | Experienced in AI product development | Patented digital biomarkers

Dr. Nilakash (aka Neel) Das is a clinical data scientist, whose work lies at the intersection of artificial intelligence and respiratory medicine. As a budding researcher, he has co-authored more than 15 scientific publications. He is also an inventor of two patented algorithms for respiratory diagnostics, and is a recipient of several clinical science awards. Neel completed his bachelors from the Indian Institute of Technology, Kharagpur and his masters from Technical University Delft in the Netherlands. Presently, he works as a post-doctoral researcher at the laboratory of respiratory medicine and thoracic surgery at KU Leuven, Belgium. He also collaborates extensively with ArtiQ, a spin-off company of their laboratory, in translating AI research into clinical practice

In another lifetime, almost six years ago, I was designing oil platforms and vessels. Little did I imagine then that my life was about change its course. I could feel the drift, my conscience plunging into an internal turmoil that the work I did was irreconcilable with the global havoc wrecked by climate change. Therefore, I just jumped ship after spending almost a decade in that field.

Around the same time, the real potential of artificial intelligence (AI) was starting to unravel. Rapid developments in deep learning, which had burst into public consciousness (Alexnet by Krizhevsky et al. 2013), signalled the dawn of a new era.  Algorithms were now capable of automating tasks that once required the epitome of human perception.  In fact, I even tried some of my half-baked AI ideas in master’s thesis not to the amusement of my supervisors who were classical physicists.

I knew in my heart of hearts that a doctoral apprenticeship in applied AI in medicine was the perfect opportunity to learn and to make a meaningful contribution to improve human lives.  It would have allowed me to take a deep dive into the exciting world of clinical data science. These were good thoughts to have. But why would a field as conservative as medicine would accept me when I have ‘oil’ on my hands?

In the summer of 2016, my co-promotor had advertised a position that he was looking for a Ph.D. candidate to build mathematical models on spirometry (a medical test to monitor respiratory health). I had no idea what spirometry was, but that did not stop me. That was the only position I had applied that summer, and I was overjoyed when I received a call for an interview.

I clearly knew that there were candidates that were far more qualified. To ‘sell’ my engineering skills, I came prepared with a few slides on how to fit a non-linear model on observed lung pressure-volume loops (I had no idea then on the physiological phenomenon!). I vividly remember that joyful moment when I received an email congratulating me! It had come after a long time of stressing about my future, and my dwindling finances. Years later, my co-promotor confided to me that he did not consider my CV initially as it was so alien to the respiratory field, only to have been convinced by a friend of his to take a second look as I had some mathematical skills. I thank that person wholeheartedly!

However, the Ph.D. journey was never a smooth ride. In the beginning, the learning curve was overwhelming. Not only I had to educate myself on classical statistics and AI, I also had to soak in an enormous body of literature on lung physiology and respiratory medicine. I am indebted to all the amazing online resources and MOOCs that made my ride much smoother. Without these resources, I shudder to think if I could have even completed my Ph.D.

Nonetheless, the real challenges arose when the time came to communicate my results. Initially, my presentations would be highly technical/mathematical which would amuse or even frustrate my colleagues, who were mostly clinicians and biologists. It was the same story with my research manuscripts that were being rejected by one journal after another. It was at those lowest moments that my supervisors played a big role in shaping my attitude. They taught me how to transcend from thinking only about the technical finer points of an algorithm, to deeply caring about its clinical aspects as well. I also attended countless presentations to understand the nuances of scientific communication. This led to a dramatic transformation in my scientific outlook, resulting in over 15 peer-reviewed articles (including Nature), 2 patents, 3 clinical science awards and several public presentations.

As I reflect on my Ph.D. journey, I feel that it was the best decision I have ever made in my life. What started out as an avenue to learn more about AI, became much more. I learnt to do science. Asking the right questions, translating a gut feeling into a hypothesis, using data to prove or disprove it, has been an exercise of enormous intellectual stimulation. It imbibed in me the scientific rigors of analysing, interpreting, arguing and communicating clearly and logically. The most important of all, it taught me to be humble about the fact that I know only an infinitesimal amount within the infinitely large corpus of knowledge. An opportunity, such as a doctoral apprenticeship, in which my contribution led to a minuscule gain in understanding human health, gives me immense pleasure.

 

To learn more about Science news and inspiring stories, Please visit  mScience Magazine at ScieceBlurb

Leave a Reply

%d bloggers like this: