Safura Suleymanova: Dancing with Data Science

In the age of disruption, the ethos is often, move quickly, pivot, break everything, pivot again. This rude energy clears a path for the new: self-driving vehicles, neurosynaptic computing, and next down, quantum. In the case of Safura Suleymanova, the discipline of data science — and her own tenacity — brought a shining new talent across the globe from Azerbaijan to Ontario to solve the problems of global industry.

But like other data scientists I’ve spoken to, Safura describes the power of her craft not as the blunt force of a sledgehammer, but as an art requiring balance and patience. Read on to hear her measured approach to thorny problems in data, and in life.

You’re originally from Azerbaijan. Tell me how you came to Canada.
I came as an international student to study mathematics — I’d wanted to ever since I was a kid — at the University of Waterloo. But pure math wasn’t as much fun as I thought it would be. I couldn't relate the concepts we studied to anything to do with real life, like business. I discovered data science and noticed, oh, here are concepts from linear algebra I can apply to real life problems and come to conclusions that help people.

How did you go from student, to data scientist at IBM?
It was a long path. In Azerbaijan, my first job was as a procurement specialist at a cement company.

So, you went to college in Canada, then back to Azerbaijan, then came back to Canada to work… how easy was that?
It wasn’t easy at all. But I did it because I had a desire to be in this field and there was no room for data science in Azerbaijan. I worked for a very big company analyzing financial statements manually, so I spent my days going through line after line of Excel. I knew there was a place for, for example, logistic regression to classify whether a person is going to be able to repay a debt or not. I applied the process and reduced the time from several weeks to two days. But even though I was able to apply it at work, there’s a huge amount of resistance to novel ideas.

How did you handle that?
I educated myself. At that point in time I didn't believe that I had enough skills, so I spent a year studying while working part time in Azerbaijan.

It was then that I found IBM’s Cognitive AI site online and took free courses in data science; that was my real introduction to the field. From that point on, I knew I’d like to work for IBM, because IBM had introduced me to my passion.

What excites you about data science?
That every problem is different; I’m never bored. I meet clients in telecom, for example, who’ll say, ‘We see you have experience in this industry, tell us how it’s done.’ I tell them, ‘I’m confident your case is going to be completely different’ — and it’s proven to be that way.

Tell me a little bit more about Azerbaijan. It’s intriguing to talk to someone from there. What do you do for fun there?
I’m from Baku, the capital city. There are new sections, where one can find ultra-modern buildings of the kind you see in Dubai, and there is an "Old City" with buildings from the 12th century, right in the city center. My idea of fun is to go to the countryside with my family, then eat out — Azerbaijani people like good food.

Where do you think you’ll be in five years?
I can't predict that even though I’m a data scientist! I do want to learn more about computational neuroscience, and to combine machine learning with other fields I’m interested in, like yoga and meditation. Yoga requires patience to improve, and I have an idea of how to use ML to help. People who do yoga aren’t typically into technology, but I’ve been talking to my instructors about an app.

Interesting! What other interests do you have outside of work?
I studied ballet and was on a path to becoming a professional dancer.

Really? That’s not easy. You have to balance right up on the end of your toes, right?
Yes.

I have no skill like that. That’s impressive. You said Azerbaijan had a low level of data science. Would you ever go back and try to bring them up to a higher level?
For now, I want to stay in Canada to learn more about putting AI into production in the right way. I think companies — even here in Canada where AI is surging — are still struggling with that. Maybe I’d go back in ten years but putting AI to work is not a job for one person. It’s teamwork, and everyone has to be ready. When I talk about AI in my country, people say, ‘Oh, AI is robots coming to take over my job.” It’s hard when there is that kind of resistance. And there’s an imbalance between men and women in the workplace, so it can be difficult to convince management to try new things.

You’re very calm. Is that the yoga?
It must be, because in this field it's very hard to be calm. Everyone sees data scientists as the answer to all their problems and all their questions. You have to work hard and gain their trust. You need to teach them that AI is not magic, and it’s not to their advantage to jump right into it just because everyone is talking about it. It’s a balance. But we’re very lucky: our clients are collaborative. They see us as part of their teams.

Hometown: Baku, Azerbaijan

Currently working on: using weather and health data collected from smart watches to predict the risk of a person catching influenza

Favorite algorithm: I don’t have one, but my favorite data science moment was when I used the Bhattacharyya distance metric to compare two groups of users based on personality traits.

Favorite yoga pose to clear my mind: child's pose (in Sanskrit: "balasana"). I like also rolling my head from side to side to give my head a little massage — nothing complicated and can be done by anyone.

Dinesh Nirmal – Vice President, IBM Analytics Development
Follow me on Twitter: @dineshknirmal