Data Scientist, Novetta, Washington DC


I started at Yale Law School strongly suspecting that I would not practice law. By the time I graduated, that suspicion had been confirmed. I interned at a United States Attorney’s Office, and worked a summer at a large law firm, but neither felt like a place where I wanted to build a career. During the first of my two judicial clerkships, I realized that I had been thinking the job hunt in the wrong way. Whatever I did, I would always have a YLS degree. What mattered was not that I “use” the degree in some way, but that I find the right job, whatever it was.

With that insight in mind, I cast a wide net, thinking seriously about my own passions, what made me happy, and what my ideal job would look like. A month or so into the job search, I had the good fortune of reconnecting with an old friend who worked as a data scientist. After speaking with him for a few hours about his work, I realized the job had much of what I wanted: mathematical problem-solving, computer programming, rigorous approaches to practical problems, and a good work-life balance. A competent data scientist must possess: familiarity with at least basic statistics, significant mathematical aptitude, knowledge of scripting languages such as Python, and knowledge of database structures and querying languages. Most data scientists have degrees in computer science, statistics, or math. While a degree—in particular a graduate degree—is not necessary to enter the profession, if you want to be a data scientist you will need some way to demonstrate your ability to do the job.

A law degree does have some value. Rigorous analytical thinking can help frame problems clearly. An understanding of the underlying legal or policy issues involved in certain projects makes it easier to build a solution of true value to a client. Most importantly, the ability to clearly and precisely explain ideas is invaluable when presenting results to sometimes-unsophisticated audiences. Still, a law degree is not a credential of much value to firms hiring data scientists. Some might view it with skepticism. Most will ignore it and focus on the hard skills required by the profession.

My day-to-day work as depends on the client and the problem. Some clients engage Novetta to build complex machine-learning models, while others are interested looking for more straightforward, easy-to-interpret analyses, and a few are more interested in helping improve their existing processes, whether by refining the enterprise architecture or testing pre-existing statistical or business models.

Some tasks are common to almost all projects. I spend a great deal of time exploring the available data, understanding how it fits together, cleaning it, extracting features, and generally putting things in order for model building. Any modeling project will involve fitting and tuning. Any project which fits into an existing IT environment has to be unit tested and properly integrated.

Because I work for a consulting firm, I move from client to client and I therefore tend to see smaller pieces of the picture. One client might be interested in building or improving a model to detect money laundering in financial transactions, while another client wants to use natural language processing to detect trends in foreign media coverage. Data scientists working in-house will focus, in most cases, on a particular subject area and become intimately familiar with particular data sets kept by their firm.

Moving from law to data science—at least form biglaw to data science—typically involves trading money for work-life balance. Most data scientists make good money, but far below the salaries typical of biglaw firms. To be concrete, an entry-level data analyst or data scientist typically makes over $60,000, in some cases over $100,000 per year. Salaries grow from there, but only the truly elite in of the profession will top $200,000. On the other hand, data scientists working in-house, and those who consult for more specialized firms, tend to work reasonable hours. I rarely work more than 40 hours in a week, and I have the flexibility to work the schedule I choose, subject to client meetings.

My hunch is that most students reading about my career path won't be especially interested in data science. If I'm wrong, and you want to follow the same path I did, then please talk to me. The field is changing rapidly, and beyond the generic skill set, it pays to keep up with the hot technologies and trends. I'll talk your ear off about that. If I'm right though, and my job sounds thoroughly unappealing, I hope you can still take something of value from my story. A YLS degree is a fantastic credential, but you should never view it as a constraint. Life is better when you can enjoy your job, even if it makes little use of your law degree. If nothing else, “I left the law to become a data scientist” is always a good conversation starter and a great story for a first date.