Associate Director, Boston
At BCG, Advocates are thoughtful and engaged. They listen closely because they believe that the best solutions can be discovered in the most unexpected places. Arun challenges the status quo.
The best part of working at BCG GAMMA is the family atmosphere. People watch out for you, want you to succeed, and help you get there. You are not treated like an entry in some Excel file.
My background is in machine learning and deep learning. During my PhD, I studied the underlying math that drives deep learning models and now, thanks to advances in computing speed and big data technology, we can bring these deep learning models to life and benefit from them. At BCG, I build and deploy machine learning, deep learning, and artificial intelligence models in various industries, including consumer goods, pharma, and oil and gas.
BCG has been a great place to use and hone my skills. BCG GAMMA has a symbiotic relationship with the BCG generalist team, whose members help define analytical use cases, success criteria, and deployment strategies—while we focus on data science and data engineering. It’s a healthy marriage, where we push each other to bring our best. The partner team always pushes us to ensure that we made the right assumptions on inputs and outputs, and we document limitations of our models in detail. We bring the scientific rigor, and they bring the business rigor!
Arun holds a PhD in cognitive and neural systems (computational neuroscience) from Boston University, and a bachelor’s (honors) and master’s degree in electrical engineering.
Q: What do you like most about your job?
A: Endless creative opportunities. In the last couple of years, most of the cases I have worked on are extremely innovative from a scientific perspective. Based on the client needs and data, we adapt existing algorithms and/or create new algorithms (and file patents on them). I love translating the business problem into scientific formulations and proving the merit of our innovative approach by showing tangible business impact. Explaining these complex algorithms to C-suite executives is quite a challenging job that I enjoy the most! In my years in graduate school, we had this saying that if you cannot explain your research to your grandmother, you don’t know what you are doing. Luckily, most of my clients are easier than my grandmother!
Q: What characteristics do you believe define a BCGer?
A: High standards. Everyone has high standards for themselves, which collectively translates to higher standards as a team. BCGers are also very humble about their high standards and smartness—since they know their colleagues are smart in their own way. When you compete with yourself, you get the most gratification!
Q: How do you maintain a sustainable career balance at BCG?
A: BCG is extremely brand conscious, and we don’t overpromise. We ensure we can deliver what we promise. I recall being in a client pitch meeting within the first few months of joining BCG, and the client asked if I could build a model and improve performance by a factor of 3. I replied that we could do a factor of 1.5, since the data did not capture a lot of exogenous drivers. The senior partner told that client that we would be walking away from this opportunity. This is very different from other companies, where if the client asks for a 3x improvement, you are required to promise more than a 5x improvement, and you bank on some junior data scientist to figure it out!
Q: What are some of your favorite things to do outside of the office?
A: I am a long-distance runner. For the last few years, I’ve done at least one marathon a year. I also enjoy time with my family, and I support a bunch of small theater companies where there is a lot of talent that needs to be nurtured.
Q: If you weren't working at BCG, what would you be doing?
A: I’d be doing improv comedy shows in the evening and probably teaching machine learning or probability theory in the daytime. I would probably move close to the mountains, so that I can hike more and volunteer more for mountain rescue operations. Search and rescue depends a lot on probability theory!