Sandip De

Born in India, where he studied physics in Kolkata and Pune, Sandip De moved to the University of Basel for his PhD, where he worked from 2009 to 2012 with Stefan Goedecker on nanoclusters and global geometry optimization problems. He subsequently joined IBM research in India as a staff research scientist and worked on semiconductor research and development for almost 3 years. He then decided to go back to Switzerland and dive into machine learning methods for accelerating materials modelling, an interest that led him to join NCCR MARVEL and the group of Michele Ceriotti at EPFL in 2015. In 2018, after leaving EPFL, he joined BASF SE in Ludwigshafen and now leads a global team working on quantum mechanical simulation for solid-state materials and inorganic chemistry.

Interview by Nicola Nosengo, NCCR MARVEL, September 2023

What was your role in NCCR MARVEL, and what did you focus on at the time?

My project was about developing novel machine-learning algorithms fundamentally critical for application areas encompassing a diverse set of materials, including inorganic materials and molecular crystals. The goal was to do fundamental method development in supervised and unsupervised machine learning related to chemical and physical property prediction.

What was your path after leaving MARVEL?

I was very clear about the kind of position I wanted. I wanted to go back to industry and work closely on applications-oriented problems, that strongly align with the present challenges our society is facing. During my previous experience at IBM, I realized that I enjoyed working on practical solutions that can have a tangible near term impact beyond just writing academic papers. I started to look for positions quite early on, about one and a half years after joining my postdoc. It’s challenging to find a core research job in the industry, where you can do fundamental research and still get to work on industrial problems. I was very picky about that and ready to move anywhere in the world to do that. In the end, I had three opportunities: one in Canada, one in London and this one in Ludwigshafen. I chose the BASF job, as I found it to be the most relevant for my field of expertise, highly exciting and potentially most impactful. Equipped with all the skills I had gathered until then, I was ready to apply machine learning and fundamental materials modelling to the transformation of chemical processes. What could be a better place to do that than one of the biggest chemical companies?  After all,  you are sitting between the experimental data and the big chemical plants, where a small improvement could make a big difference not only for the company but for our society.

How can you describe your current job?

I initially joined the quantum chemistry modelling group, and my focus was mostly on heterogeneous catalysis:  modelling complex interaction of molecules and catalysts and understanding how they affect the catalytic process. It was very similar to the work many people do in MARVEL, in atomistic modelling. I have aimed to accelerate the process using machine learning and big data-driven workflows which are critical to meet tight industrial timelines goals.

Since last year I have been in charge of leading the team working on quantum chemistry for inorganic materials. It’s a global team, with people here in Germany, in the US, in China and in India. My team works on inorganic chemistry, including electrocatalysis, battery materials and of course heterogeneous catalysis. BASF has very ambitious sustainability goals, a 25% reduction in emissions by 2030 and becoming net-zero by 2050. Catalysts play a key role in new sustainable processes. We work on cutting-edge technology for new catalyst development that is very high-risk-high-gain by nature. I also lead several academic collaborations, co-guide PhD students and postdocs, and take part in publicly funded activities to continuously build our in-house research capabilities. These digital capabilities combined with the expertise of our experimental colleagues help BASF move towards its sustainability goals.

How did the MARVEL experience prepare you for this position?

From a scientific point of view, a major part of my technical strengths and approaches is an evolution from what I learned during my MARVEL times. The underlying motivation is still using machine learning methods to accelerate materials modelling. This is as yet a rapidly developing field quite far from maturity and It is still not common practice to apply these methods in industry. My background and connections give me the advantage of being a super early adopter of these new technologies for industrial applications. On the other side, MARVEL is a unique program where there are several meetings and conferences where you get to interact with the public and funding agencies. I remember that people from the SNSF would come and talk with not just professors but also with postdocs and PhDs. The program encouraged different academic groups to interact and work together. This interdisciplinary experience was quite crucial for my development. I was also involved in organising MARVEL retreats, and I met my present boss at BASF for the first time at a MARVEL retreat! I also got to learn skills about how to communicate with the general public and how to work with an editorial news team, when my work was highlighted by EPFL news, and looking back now, I think those skills have been equally important in my career in the industry. 

What advice would you have for young researchers who are considering a career in industry?

The transition from academia to industry can be motivated by various factors. First, let me address the group that wants to stay close to research.  Early in a research career, sometimes, it may stem from a sense of frustration, where individuals doubt their success in an academic career. My primary advice is for individuals to introspect and identify what truly brings them joy and fulfilment. A Scientific career irrespective of academia or industry is bound to go through these phases. You must be able to enjoy the journey. Understanding your motivations and strengths is crucial before making any decisions. Additionally, evaluating the practical limitations, such as familial obligations that might influence your choices, is essential. If you are committed to working on topics of your interest, know that you are doing outstanding work and are flexible enough to move to any place that offers it, there is no reason to compromise on that research career dream! Stick to your academic career goals.  An industrial research position can give you a deep sense of fulfilment as well when you see you are playing a part in solving a real-world problem. However, I must admit, that such positions are limited in number as well. However, In today’s world, there are increasingly more examples of people switching between industry and academia (both ways) at both the early and late stages of their careers. Keep striving to do meaningful work irrespective of where you are.

One caveat you must be aware of in the previous suggestion is that simultaneously,  you need to cultivate a pragmatic and open mindset. It is generally not a good idea to continue working as a postdoc for years in the same group you got your PhD from or just jump from one postdoc to another for years without thinking of long-term goals in your research career.  Awareness of what research options are available outside academia is also a challenge and is often full of outdated prejudices (I know, I had!). If you are fortunate enough to work on topics which have industrial demands, why not take an early opportunity of internships or postdocs to get first-hand experience if you don’t get a permanent position immediately? you will only strengthen your skills by doing so even if you decide to go back to academia.

On the other hand, if you realize a scientific core research career is not for you and you want to explore other adventures, in the business world there are many opportunities, and there are indeed skills that you learn while doing research that can be transferred to other positions. Data science, machine learning, and programming are practical skills that are very much in demand now in a wide array of jobs. Pick up those skills irrespective of what you plan to do later. Start exploring what type of jobs excite you and start preparing for them as soon as possible. Often the potential options are higher right after PhD if you want to make a complete switch.