Sattwati Kundu is a Data Scientist at IBM, India Software Labs since 2018. In her current role, she is responsible for contributing to innovations in the areas of applied analytics in the Internet of Things. She carries with her 14 years of rich industry experience in the areas of analytics and predictive modelling in the field of Oil and Gas Upstream businesses, working with Shell, Schlumberger and Jubilant Group and in the field of Information technology, working with IBM. She also led the productization of analytical components in a start-up for a short period of one 1year before joining IBM. 

Sattwati carries with her in-depth knowledge in the areas of analytics and integrating information gathered from different sensors deployed in the oil wells. In her past tenures in the Petroleum Industry, she was responsible for analysing data from these sensors to build integrated predictive subsurface models. She also has expertise in the area of Wells, Reservoir and Field maintenance which involves translating various signals into meaningful actionable insights based on mathematical models. 

She has been awarded several awards during her tenure in Shell and at IBM. Sattwati holds multiple patents in the field of analytics which are in filed and to be filed stages.

Sattwati, can you tell us about your AI journey?

As a graduate from IIT KGP in Geological Sciences, I have been deeply involved in data analytics and predictive modelling since she passed out from IIT in 2006.

I started my journey in AI in 2006 when I joined the oil and gas industry as a Petrophysicist (A physicist specialized to work on “Petro”, ie., Rock). However, the definition of AI or data science was different in that industry then. My experiences from then include building a predictive model to generate quantifiable insights from data gathered from multiple sensors. The methods involved techniques ranging from advanced statistics to simple neural nets. 

Nevertheless, that had been the paradigm for Oil and Gas Industry then, which is way different from present-day AI which involves using more Deep Neural Nets, especially in areas of images and language processing. 

While at IBM, I have been responsible not only for innovating and delivering AI solutions as a hands-on data scientist in IBM IoT offering IBM-TRIRIGA but also for leading a global team of data scientists across various other IoT offerings. Additionally, I have also led multiple AI evangelizing efforts across IBM and the Labs. I have also participated in IBM’s community initiative like contributing to the designing of the IBM-CBSE AI curriculum.

What are the major challenges you faced as a woman in reaching where you are right now?

In the field of analytics, the job often goes beyond regular hours because of its research nature. Being a woman, who has several other personal commitments and responsibilities, sometimes meeting targets at both places- office and home, has been a constant challenge. 

In the Oil and Gas industry, my job also demanded me to move temporarily (few times) and permanently to obscure locations. It posed serious challenges to my personal life especially as a mother of a young child. Although it meant being comfortable with my job but very uncomfortable with my personal life, I decided to move out of Oil and Gas.

The move into AI was not only about regaining balance on the personal front, but also furthering my quest for learning and doing more in the field of AI. It was my wish to move out of my comfort zone and was very challenging as I had to learn a lot more than what I was comfortable with. I had to invest more hours in my upskilling, learning the new industry I had joined, and apply my knowledge from past experiences in the field of AI to start innovating. At IBM, the flexibility has helped bring out the best in me and strike a balance both professionally as well as personally.

What made you interested in AI?

As I had already been in the field of AI, with the latest advancements in the field of Deep Learning, I wanted to explore more than Oil and Gas data and methods. That's why I stepped out of the Oil industry in 2016 to expand my horizon across industries.

What is your area of expertise in AI and why chose that one?

I possess expertise in Sensor Data Analytics. Ranging from white-box methods, which are more governed by the physics of the system, to black-box techniques which are more data-driven; I have been working with the entire spectrum. Hence, I can understand what the solution requires to address a problem and design it accordingly for better outcomes. Sometimes this results in innovations that qualify for patentability.

What's the one thing that you see AI transforming completely? 

More than things like self-driven cars or robots, an area I believe AI will be completely transforming is our day-to-day life. For example, look at things like face-ID on phones; ability to run machines in an automated way- like cooling systems in buildings based on sensor analytics; fastest route predictions; most relevant product reviews on a website, and more

What is your biggest AI nightmare?

When machines prediction goes wrong- that is my biggest AI nightmare. As an AI creator, I know how AI works and when it can go wrong. But because it is entirely data-driven, sometimes wrong predictions cannot be avoided in runtime (i.e., while people consume the results). However, they can be rectified in post-analysis and hence it is important to check for drift and bias in AI models and thus bring trust and transparency in AI. 

What is your advice for other women who wants to pursue a similar journey?

Believe in your work and voice it. Keep learning. Upskilling to be at par is the key to success. There is no shortcut to success, failing is natural, so embrace your failure and come out stronger. Never give up!