I work as a data scientist at a bulge-bracket bank. I joined not long after getting my Ph.D. in mathematics from an Ivy League university and attending an NYC Data Science Academy boot camp. I’ve been working here less than a year, but I have to say, it’s a pretty sweet gig. My team is built like a startup within the big bank – we have our own separate floor, so it’s like a bank trying to launch a startup. I feel free to share my ideas, so the transition from academia was less pronounced than you might think.
To be sure, there are downsides that go hand-in-hand with working within a huge bank, problems that I didn’t have before as a student. To get certain things done takes a lot of time – there are various hurdles to putting things into production and working within the greater infrastructure can be challenging, even just to update our digital products, that is, getting things to work within a bigger structure in terms of compatibility.
On the flipside, I have access to a huge amount of interesting data and a ton of resources that I wouldn’t otherwise have, so I can’t complain. Big data is thrown around as a buzzword at a lot of companies, but that is what we live and breathe.
Big data IQ
The data science team I work on is currently 20 people, and we’re looking to grow over the next year.
Our job is really to leverage all of the data that the bank has to better understand who our customers are so we can better personalize their experience and better market to them. As you can imagine, the role involves analyzing data, reporting to other members in the firm and partnering with other teams, basically looking into the data, seeing what’s out there, crunching numbers and presenting our findings for the use of other people across the bank.
Almost everybody on the team has a Ph.D. in a STEM field, primarily math and computer science, but we also have a bunch of physicists, engineers and people who majored in statistics, as well as one who was a double-major in aeronautical engineering and computer science.
Data science boot camp
While I really enjoyed my Ph.D. program, at a certain point I decided that academia wasn’t for me, and I was lacking certain skills for the type of data science job I was looking for. So, I decided to participate in the NYC Data Science Academy boot camp, a 12-week program that’s lecture-based, where you’re learning programming, statistics and filling in the gaps for people like me, whereas for others there was lots of new stuff. In addition to the lectures, we did a bunch of projects to apply the skills that we were learning and build a portfolio to show off to prospective employers.
Ph.D. research skills are important in data science jobs, which is why our team has a lot of Ph.Ds on it, but there is additional knowledge that you need to know. In addition to statistics and programming, you need business know-how to properly communicate what you’re doing to people who aren’t computer scientists – that’s where the boot camp stepped in. You actually use the skills you learn to do some projects like what you would do in a job.
For example, you use the basics like Linux, Git and SQL. The programming languages R and Python were covered in the boot camp, and we used them to create machine-learning algorithms, models and applications.
The bank uses both of those, Python more than R. We also use big data tools such as Spark, Hadoop and Hive, which the boot camp also covered. Our data scientists have to be fluent in statistics and regression of different kinds, and at least be familiar with deep learning and neural networks.
Tell me about it in plain English
A key part of the job is being able to explain what you’re doing to non-technical people working at the bank. Being able to explain your analysis is as important as actually doing the analysis – how are we going to use it and explain a super-complicated model to someone in another department?
Projects I did in the boot camp were not directly related to financial services, but some were not so far off. For example, one project I did was roughly predicting Airbnb bookings, whether or not someone will book, and what ads you are more likely to click on. Our team has worked on similar problems on the financial side of things.
The trading and investment side is a bit different – but not that different. There are certain financial skills and tools needed for predicting prices and stuff like that, but the base knowledge is similar to what I have. People I graduated with who had similar areas of focus went into quantitative trading.
Advice to aspiring data scientists
Between my Ph.D. program and the data science boot camp, I felt prepared on day one, but there was definitely training on the job too. Every company is going to have their particular systems and tools that they are going to use, but in terms of baseline knowledge, I was pretty comfortable when I started.
For me, the NYC Data Science Academy boot camp was an excellent experience. I did have some gaps in my schools, so if I were to have applied for jobs right after I graduated, I would’ve had a tough time or had to take a job I wanted less. After the boot camp, I was very happy where I ended up. My current job came through a connection at the boot camp. I got an interview and got a job offer very soon after the boot camp ended, so both the knowledge I got and the hiring process was very good.
Something that’s popped up recently is Master’s programs in data science. That was not a thing five years ago. I enjoyed my doctoral program, but I wouldn’t recommend that someone to do a Ph.D. in something they aren’t interested in, whether it’s math or physics, and expect that your interest in data science will evolve over time. For undergraduates wanting to work in data science at a bank, a Master’s program is a good option. Learn as much math and statistics as you can, get a job at a startup or a bank and build your career from there.
The author contributed this piece on condition of anonymity. Bob Piper is a pseudonym.
Photo credit: Ivan Mikhaylov/GettyImages