DSI Alumni Download: Q&A with Pranjal Bajaj, Data Scientist at Boston Consulting Group GAMMA

After graduating with his master’s degree from the Data Science Institute at Columbia University (DSI) in May 2019, Pranjal Bajaj was hired as a data scientist at Boston Consulting Group (BCG) GAMMA.

BCG is a management consulting firm with more than 90 offices in 50 countries. GAMMA is its data science and AI consulting team through which data scientists, software engineers, and other specialized analytics experts provide businesses with data-driven solutions.

Bajaj completed undergraduate studies in economics and law at the University of Cambridge, England. During his master’s program at DSI, he also undertook the University of Chicago’s Data Science for Social Good Fellowship to help the Portuguese government design a machine learning platform to combat long-term unemployment.

Here, he discusses his work at BCG Gamma, which uses data science to help businesses solve some of their toughest challenges, and his understanding of how businesses and organizations use data to create positive impact.


Describe a typical day in your role with BCG Gamma.

My role varies depending on the kind of project and industry, but at its foundation lies strategic thinking and analytics, which are used to frame business problems and drive solutions. For example, reducing waste via inventory optimization and accurately forecasting demand or improving customer experience via personalization. Teams can vary from sizes of four to six to more than 30 while projects last from as little as four weeks to one year or more. We often work in interdisciplinary teams consisting of generalist BCG consultants as well. It’s a dream job. I get exposed to a variety of projects with a multitude of clients by leveraging machine learning and advanced analytics to create positive impact.

On which kinds of projects do you work?

Our projects can be bucketed into three types: demand forecasting, optimization, and customer journey or life-cycle, which also includes predicting churn and personalization. BCG attracts clients from all sorts of industries–retail, technology, transportation, social impact, health care and so on.

Which data science skills do you use at work? 

The skills I use vary from project to project, and within a project they vary depending on what stage it is in. At the beginning of a project, I may spend time setting up the infrastructure such as the cloud, databases, and GitHub repositories. Next, I will explore and visualize various datasets we receive from the client to gain a deeper look into the business problem and test various hypotheses. The following phase usually involves building models and choosing evaluation metrics to meet the defined business objectives. The modelling aspect can involve building a time-series model, a recommendation engine, or a causal model. Most of my work so far has been in Python, but the tools change based on the problem at hand. I may also expect to work in Spark or R depending on the project.

Can you discuss a project you completed recently? 

I recently worked with a large technology company to help them forecast the daily volume of customer complaints across a few products so that an appropriate headcount of engineers can be designated ahead of time to deal with these complaints. The challenge lied in appropriately identifying why pre-existing forecasting models failed to yield to correct forecasts, but the solution to this problem lies beyond what algorithms could address. It involved changing not only how the client was forecasting, but what the client was forecasting and combining both BCG’s advanced analytics capabilities and core consulting toolkit to drive positive impact.

Which aspects of your studies at DSI help you most in your work?

The core programming and machine learning toolkit have been crucial. Classes that have been especially useful are Algorithms for Data Science, which exposed me to computational thinking and optimization; Introduction to Databases, which helped me develop my data engineering skills; and Machine Learning and Applied Machine Learning, which have been my bread and butter skills at work. Additionally, my semester working as a teaching assistant for professor Andreas Mueller’s class, Applied Machine Learning, stands out remarkably since it really solidified my machine learning intuition and knowledge of scikit-learn.

Do you have any advice for current DSI students?

Build your network and use it to learn about opportunities in data science and land the ‘right’ job for yourself. DSI sets you up for some of the most competitive jobs in industry, but getting there will require mettle on your end. Given how fluid definitions of the term ‘data scientist’ are across industries, firms, and sometimes even within firms, it is important to be well aware of these distinctions.

Learn about the different employers, the kind of work they do, what’s expected out of data scientists on a day-to-day basis, and where their data scientists end up after three to five years on the job. The best way to learn about this is to actively build your network and leverage it to learn this and more. Don’t just add people on LinkedIn for the sake of it. Get them on the phone or meet them for coffee to learn about what they’re doing and get them to refer you.


Media Contact: Robert Florida | rsf8@colmbia.edu | 201 725-6435, mobile


550 West 120th Street, Northwest Corner Building, Suite 1401, New York, N.Y. 10027    212.854.5660
©2020 Columbia University