Data Science at Jellyfish


Di Wu


Could you briefly introduce what is it like to work in Jellyfish’s data analytics team?

Jellyfish delivers digital marketing solutions across the world with award-winning combinations of technology and talent. We offer paid media, data analytics, UX, SEO, creative, development, and training to support clients in a wide range of industries. So we get to use data to solve marketing problems in a variety of fields – such as education, retail, financial services, technology, healthcare, travel, media, or entertainment. And since we’re a global company, working with our data science team means collaborating with different colleagues and clients to tackle unique challenges all across the world.

At Jellyfish, data is at the center of everything we do. Our data science team builds cutting-edge solutions to help our stakeholders and clients to make better decisions, improve operational efficiency, and increase profitability. There’s always a new set of data science challenges, so new ideas are encouraged. Our team gets plenty of opportunity to innovate. On a given day, we might wrangle large structured and unstructured datasets, model and parse for specific text and image elements, apply machine learning algorithms to discover hidden patterns and customer insights, or build prototypes for production-ready systems.

We understand that you worked with NASA as a research scientist prior to Jellyfish. What are the differences between these roles?

My previous work was more research-intensive. Research work tends to be highly specialized, and involves going to great depth to examine how a theoretical concept works. As a researcher, you are responsible for producing ideas and publications to procure grants that follow a longer timeline.

In contrast, my current career in data science is very much practice-based, with a focus on building models that yield quick, concrete results. In addition, as a data scientist, you often bridge the technical findings of your models to commercial considerations as an initial input. So I’ve found that data science means more frequent and varied projects that are fast-paced with shorter timelines in mind.

Do you feel like your research experience helps you to fit in the role you are now?

Yes, definitely. Often in data science, hypotheses need to be tested and validated with the right amount of data. So the structured mindset and curiosity I developed from my research experience has provided me a well-rounded framework to dissect relevant information and garner further insights.

My research experience also instilled strong communication skills that were necessary for participation with numerous publications and presentations. This set a good foundation for the business world, where I’m constantly presenting the results of my models to different audiences.

How does someone stand out in data science?

As most data scientists come from a STEM background, technical skills alone are not enough – those are more of a prerequisite. So the best way to stand out is to display a high level of creativity.

Being creative in data science means thinking more deeply about the domain problem at hand, and trying to tie it back to your data science toolkit. For example, in digital marketing, you need to build a deep knowledge of regulations or policy changes (e.g. GDPR within digital marketing) and translate that into the data science models you build. It is this second-order thinking that will make your data science work stand out in the eyes of managers.

What advice can you give to aspiring data scientists? How should one break into this field? Is a PHD necessary?

There are broadly two paths that an individual can take within data science – research or application focused.  For the research path, your area of responsibility will be very much centered on developing the latest cutting-edge machine learning algorithm that solves a specific research or business problem. For example, optimizing the advertising bidding algorithm that feeds into the real-time bidding production system. For this route, I would recommend getting an advanced degree such a PhD in a highly technical field.

On the business application side, your focus would be to provide actionable insights or recommendations from data, and perform various data-related tasks. These could be anything from defining metrics, to running and interpreting experiments, to creating analysis, drawing causal inferences, or building predictive models. This activity will draw from a combination of skill sets such as data wrangling, business acumen, math, and software engineering knowledge. All of these can be gained through continuing education and working experience.

What does Jellyfish look for in prospective applicants?

A curious mindset and an aptitude for research is crucial. Any prospective applicant should display a desire to understand the rationale behind how things work. In addition, candidates should have strong engineering skills and a passion to build and invent solutions.


Disclaimer:
These points reflect the personal views of our interviewee. Any of the discussion points does not reflect any of the organisations she has worked or is currently working at.


Create your website at WordPress.com
Get started