What kind of data scientist should you be?
Amazon's Daliana Liu helps others in the field chart their own paths.
[Editor's note: In June 2022, Daliana Liu accepted a role as a senior data scientist with Predibase.]
Daliana Liu’s career has touched many facets of data science. Now as a senior data scientist in Amazon’s Machine Learning (ML) Solutions Lab, she uses her experience to help young data scientists decide where they fit in this ever-expanding field.
Liu was born and educated in Dalian, China — the inspiration for her chosen name. She studied mathematics and applied math, mainly because her parents had assured her it would offer her flexibility in her career. She found the work taxing and was largely uninspired — until she stumbled into statistics.
“Statistics and applied math, you can literally apply those to every industry,” she said.
After undergrad, she was looking for something new — and a master’s degree in statistics from University of California, Irvine seemed like the right move.
Data science: a continuing education
Her first job out of graduate school was as a business intelligence data analyst at Boingo, a company that provides mobile Internet access. Liu had been there for almost a year when a recruiter from Amazon reached out to inquire about her interest in a role with the company’s retail team.
“I wasn't sure whether I had the necessary skill sets, like A/B testing and SQL, but I decided to give it a try,” she says.
She studied, got the job offer, and since then has continued to learn and progress. She transitioned from the retail team to human resources and then to her current role as a senior data scientist and machine learning engineer with the ML Solutions Lab.
If you know how to apply data science, you can work in any industry.
Today Liu supports Amazon Web Services (AWS) customers seeking to incorporate more machine learning into their businesses. In a recent project, she worked with football coaches at University of Illinois Urbana-Champaign, where she and colleagues developed a ML model that predicts the results of the team’s plays, helping significantly reduce the time coaches spend on game planning. In another project, she helped Sportradar, a real-time sports data provider, predict soccer goals using a 3D computer vision model.
But Liu says many of the skills she uses today — like programming in Python — were skills developed after graduate school. Expanding her portfolio of functional skills, she adds, is part of a career in data science.
“It’s not easy, but it’s energizing to always be learning,” she says.
For Liu, the draw to data science is all about problem solving. Every industry has data, and “if you know how to apply data science, you can work in any industry,” she says.
That’s exactly what she does now as part of ML Solutions Lab. The team consults with and support clients who are exploring ML adoption. The projects tend to require two to six weeks, which means she can take on multiple projects a year, all in different industries, most with different business models.
“Their problems are very interesting,” she says. “And building relationships with customers and seeing how ML affects their business is very rewarding.”
Helping others find the right fit
Now that she’s a senior data scientist at Amazon, Liu offers daily career advice on her LinkedIn page and pens a free newsletter, Nerdy Talk with Daliana, for early career data scientists. Many of her followers come to her with the same question: What kind of data scientist should I want to be? Modern data science is a relatively young field, and there’s not a lot of guidance about where a long-term career can lead.
Don't over engineer your data science project.— Daliana Liu (@DalianaLiu) March 31, 2021
If you can solve a problem with logistic regression, don't waste your time and money on that state-of-the-art neural network.#datascience
“I wished for more material online from people who had already done it,” she says, recalling that she felt similarly lost when starting her career journey.
So, she wrote a LinkedIn post contributing just that. In it, she offers questions to help rising data scientists decide which avenue to pursue. It’s really about the kind of work you like to do, she says.
“Do you enjoy applying a deep understanding of regression, statistical theories, model’s assumptions, statistical testing, and time series?” she asks in the post. If so, you might like a career in forecasting or strategic planning. If you’re more interested in seeing the growth of the product and understanding user behavior, she counseled, product analytics could be a good fit.
And for those like Liu, who prefer deep learning, coding, and using data to build tools, she recommends deep learning or ML engineering. Liu first explored machine learning as a personal interest. She took some classes through Amazon’s Machine Learning University, courses once offered only to Amazon employees, but now available free to the public. Eventually she started incorporating what she was learning into her work building ML models to predict customer behaviors. Digging into a new interest gave Liu an idea of how she wanted her career to develop.
Liu says the response to the post, which has drawn hundreds of positive reactions on LinkedIn, was appreciative. “A lot of people have said they found the post helpful, that it inspired them to get started instead of feeling hesitant,” she says.
She also advises those starting out in data science to be vocal about interests and the direction you want to go. “There may be opportunities you’re missing out on simply because your managers don’t know you want to help,” she says.
Even though she’s made a fulfilling career for herself, Liu is quick to qualify that the process didn’t happen overnight. There were rejections along the way.
“I got rejected because I didn't have the specific skill they were looking for,” she said. “You just need to keep looking until you find one position that is a good fit for you."
She also advises those who want to explore data science to follow the Amazon leadership principle of “learn and be curious”. There isn’t a data scientist who has mastered every skill in the field, she says; the key: let your work guide you toward new skills and always be ready to learn.
“Pay attention to what makes you feel excited,” she says. “That’s key to a fulfilling career.”