After she graduates this month from the master’s program at the Data Science Institute, Aishwarya Srinivasan will begin working at IBM as a data scientist. She'll join IBM’s Data Science and Advanced Analytics team, where she'll use data-science techniques to solve some of the company’s business challenges.
While a student at DSI, Srinivasan had worked and excelled as a summer intern on IBM’s Data Science Elite Team. On her first day of the internship, she asked her manager what project he would like her to work on. He just smiled at her, she recalls, and said, “Do as you wish. Pick a project that interests you and try to build a product to help our clients.”
At the time, Srinivasan was interested in a type of machine learning known as reinforcement learning so she proposed using it to build a trading and portfolio-management model. Her manager agreed, and it turned out to be a good decision: The model she built was so successful that IBM filed a patent for it (Machine Trading Using Reinforcement Learning, Sept. 21, 2018), an outcome that delighted Srinivasan.
“During my previous internships,” she says, “I was always assigned to specific projects by managers who gave explicit instructions on how to do the projects. So when my IBM manager told me to ‘Do as you wish,’ I knew the internship was going to be a completely different experience. It was an amazing internship and helped me get hired as a full-time data scientist.”
Srinivasan had similarly amazing experiences at DSI. She worked as a graduate research assistant for DSI as well as for the Columbia Business School and Columbia University Irving Medical Center. For her Capstone project, she was part of a team that developed a model to assess medical companies based on their clinical trials. She volunteered at Datakind, a nonprofit where data scientists try to solve societal problems; writes a blog that covers data science; has several research publications; and has spoken at major conferences. As a result of her myriad achievements, she was also shortlisted for the 2019 the Women in IT Awards, in the category of Young Leader of the Year. The awards celebrate women in IT and seek to identify new women role models in the industry.
In this Q&A, Srinivasan talks about the educational and research experiences she had at DSI as well as her work for IBM.
You were an intern but you were named on a patent. How do you feel about that?
My goal throughout my IBM internship was to pursue learning and innovation. I was excited to work on a novel reinforcement learning project, which was to build a sophisticated system to perform machine trading, essentially a way of executing large trade orders using pre-programmed and automated methods. Getting a patent was just a by-product of our team’s efforts. But it definitely is motivation for me to keep striving to gain knowledge.
What will your main role be in your full-time job at IBM?
As a data scientist on the advanced analytics-data science elite team, my work will include but not be limited to interacting with clients on business challenges and using advanced machine-learning techniques to come up with innovative business solutions. Working on the team will give me an opportunity for extensive research, client presentations and conference exhibits. I’ll also get to interact with international research groups.
What did you do as a Graduate Research Assistant at Columbia University Irving Medical Center?
I helped develop a model to predict acute kidney injury. The injury is a common postoperative event, which can be an effect of the drugs given to the patient before, during or after surgery. My study analyzed the electronic health records of kidney patients as well as their pre- and intraoperative blood-flow data, from which we created a model for predicting acute kidney injury. The doctors I worked for were not data scientists, so I had think like a doctor and keep the model easy to use and explain, which was good experience.
You were also a Graduate Research Assistant for Professor Andreas Mueller, helping him with scikit-Learn. Explain that project.
Scikit-learn, the Python machine learning library, contains evaluation metrics and tools for implementing machine learning workflows. The goal of this project was to analyze the current usage of scikit-learn on a large scale (i.e., the scale of all open-source code, even all public code), and extend the library based on the findings. I helped to identify usage patterns, problematic use cases, and ways to improve the interface. Andreas Mueller is a leading expert in the field and working for him was a great experience and a professional milestone in my life.
You’ve had a number of internships before you came to DSI. Can you discuss a few of them?
I have previously interned at several organizations including Microsoft, Ernst & Young, Tata Consultancy Services and the National Informatics Centre. My internship projects at all these institutions were data science and machine learning use-cases, and I worked in a variety of industrial sectors ranging from finance, technology, government, and auditing and risk strategies. Indeed, my internships gave me a better understanding into how to leverage machine-learning techniques to render business insights and solutions.
Can you talk about your volunteer work for DataKind?
The DataDive event, sponsored by DataKind, is a meetup for nonprofit organizations that present their business challenges to data science enthusiasts and experts who participate in the event as volunteers. The aim is to leverage data science methods for social good. I represented IBM at these events and tried to solve the various challenges faced by these NGOs. They do good work to help society so it was really enjoyable to help them succeed.
Where did you study as an undergraduate?
At Vellore Institute of Technology in India, where I studied technology with a concentration in computer science and engineering. Considering my keen interest in these data-intensive fields, DSI was a good fit for me for graduate school.
Why did you come to DSI for a master’s degree?
I came here right out of undergraduate school. I did some research and saw that DSI had great professors like Professor Mueller, from whom I could learn machine learning, and Alexandr Andoni, an expert in the algorithmic foundations of massive data. I also saw that DSI gave students the opportunity to work on research, and that interested me.
How’d you first take an interest in data science?
I did some data science as an undergrad and was exposed to some coursework. My university had guest lecturers from all over the world who talked about data science and all you can do with it. After hearing them, I read a lot of research papers about big data and cloud computing and how data science was flourishing as a new field. I was very intrigued and wanted to be a part of it.
What was the highlight of DSI for you?
Working for Professor Mueller was certainly a highlight, in that it changed my life. But I’d say DSI offered me a wealth of opportunities to do research at different schools and to assist brilliant professors. The classes were great and DSI also helped me get the internship at IBM that led directly to my full time job, which I am really looking forward to.
Once you start your IBM job, you said you intend to help your parents financially.
My parents, who live in India, always gave me freedom and encouraged me to do whatever I wanted, with no gender bias. I’m an only child and my father is retired and my mom will retire soon, so I’d like to help them buy a nice house in a good neighborhood. They also helped me with tuition at Columbia, so I want to repay their generosity.