The world’s leading publication for data science, AI, and ML professionals.

It’s About Time We Broke Up Data Science

Why We Must Decentralize Data Science and How It Might Look

Opinion

Photo by Manolo Chrétien on Unsplash
Photo by Manolo Chrétien on Unsplash

Introduction

It’s highly unlikely that business owners are going to read this and begin to change their perspectives on how we define Data Science. Not because I doubt my influence or anything, but since I’m aware that the majority of my readers are at the beginning of their Data Science journey – I really dislike the term "aspiring" – but here is what I wish to tell you all…

Stop trying to be good at everything in Data Science, and pick 1 (max 2) area’s you want to specialize in and get really good at it!

Overview

Let’s face it… Breaking into Data Science is difficult for a number of reasons. However, I’ve come to a realization recently that much of the difficulty lies in the fact that the term "Data Scientist" encompasses so many different technical qualities that make it virtually impossible for one individual to meet all these criteria and stay up to date in each area – and that’s okay!

Getting A Data Science Job is Harder Than Ever

I’ve been listening and speaking to Vin Vashishta, Chief Data Scientist and LinkedIn Top Voice 2019, and he believes that for roles to be defined better then more specialization amongst practitioners must occur. I actually like this idea because it requires practitioners to invest more time in to strategically planning out how they want to put themselves forward to the marketplace and how they believe they could add value.

Much of the work we do as Data Scientists requires we take time to go deeper into a problem, which could be tough for someone who’s working on a Conversational AI project in January and a Computer Vision task the next month. Therefore, deciding to specialize is not only beneficial to the company, but it’s also better for you on your journey in Data Science since you’ll have more time to focus on tasks specific to the area you’ve focused on – and you can go deep.

"Data Scientist is a category rather than a job title. It’s really 8+ different jobs hiding under a trench coat" – Vin Vashishta

In reality, the goal is to set yourself apart from the herd which is generally done by putting emphasis on your individual special skills. What makes you, you? We all have something that differentiates us from the next person and what makes an interesting candidate is one that could emphasize this concurrently with their data skills. Whether it be your ability to do NLP tasks, Computer Vision tasks, or you’re a maestro at Fraud detection problems, we all have something that we could do really well and a compelling background story to back it up.

A company that is hiring for a Machine Learning Engineer would often have a very specific problem, such as deploying a model and serving it via a REST API. As a result of some of the most popular courses for Data Science (and other reasons), the majority of Data Scientists would have a dash of general Data Science skills so it’s really easy to put ourselves out there to be a jack-of-all-trades. To be honest, I’m guilty here.

The result of this is typically an untailored resume… An example of this is when we click "easy apply" on LinkedIn or Indeed. And it’s crazy because I genuinely get the thought behind it – "I know they want an ML Engineer but maybe they’d like that I could build a dashboard too and a whole load of other things, right?". Well… Not quite.

I’m not saying it’s impossible that you’d get a look in for this job, but from a hiring managers perspective, they’d want to have confidence in your ability to deploy a Machine Learning model and serve it through a REST API as that is their immediate problem, so it would have been better if you simply emphasized that skill.

Here’s an article that I believe delves deeper into this idea of Specializing:

Why Every Data Scientist Needs to Specialize

The Break-Up

Will there be a physical break up, like a board coming together and physically dismantling Data Science in Standard Oil style? I highly doubt it. However, I do believe that HR and Recruiters would crackdown on the job specifications to present much more specific requirements and titles. The jobs and a brief description of their roles may begin to look as follows:

Data Engineer/Architect

According to Glassdoor, a Data Engineer is tasked with transforming data into a format that can be easily analyzed. They do this by developing, maintaining, and testing infrastructures for data generation. Data engineers work closely with data scientists and are largely in charge of architecting solutions for data scientists that enable them to do their jobs. [Source: Glassdoor]

Machine Learning Engineer/Architect

In an article titled Machine Learning Engineer vs Data Scientists, the author describes the role of a Machine Learning Engineer/Architect as follows:

Machine learning engineers sit at the intersection of software engineering and data science. They leverage big data tools and programming frameworks to ensure that the raw data gathered from data pipelines are redefined as data science models that are ready to scale as needed.

Machine learning engineers feed data into models defined by data scientists. They’re also responsible for taking theoretical data science models and helping scale them out to production-level models that can handle terabytes of real-time data. [Source: Machine Learning Engineer vs Data Scientist]

NLP Engineer

As an NLP Engineer, you’d be responsible for processing and analyzing the intersection between everyday natural language and the ability of the computer to derive actionable insights from unstructured natural language data.

CV Engineer

A write up from Hired described Computer Vision engineers as follows:

As a computer vision engineer, you are able to automate various functions that the human visual system can do. You can multitask and work efficiently in a collaborative setting on critical projects to get them done. Our computer vision engineers are self-motivated and display leadership qualities.

Machine Learning Researcher

A Machine Learning Researcher has much fewer software engineering skills than a Machine Learning engineer, although they are still concerned with the advancement of a specific niche within Machine Learning. Another distinctive difference between the 2 is that ML researchers are typically educated to a Ph.D. level hence meaning they have very strong academic and research-focused backgrounds – However, an ML engineer can also have a Ph.D., but it wouldn’t be as sort after for someone that’s a researcher.

Data Analyst

TargetJobs describes the role of a Data Analyst as:

Someone who scrutinizes information using data analysis tools. The meaningful results they pull from the raw data help their employers or clients make important decisions by identifying various facts and trends.[Source: TargetJobs]

Data Analysts typically use less programming skills than what we’d call Data Scientists in this day and age. There’s often also a misconception that a Data Scientist is better than a Data Analyst – Neither is better than the other and they do different things.

Business Analyst

A business analysis constitutes for researching and detecting business needs, as well as recommending solutions to problems in business. Essentially, a Business analyst would bridge the gap between a business idea and the capabilities of the business.

Final Thoughts

Before I receive bags of abuse from the generalist community, I wish to make it clear that I am not saying that being a generalist is invaluable. To be honest, it’s great that you can perform multiple functions across the Data Science Life cycle. In the same way, a full-stack developer could perform both front-end and back-end tasks, I believe it’s pretty cool, however, if it’s hindering your ability to delve deeply into a specific type of problem then it may be better to specialize in an area of interest.

Thank you for reading! Connect with me on LinkedIn and Twitter to stay up to date with my posts about Data Science, Artificial Intelligence, and Freelancing.

Related Articles

The Specialized "Data Scientist" Will Win in The Long-Run

Tips To Land A Data Science Role in 2021

Overview Of Data Related Roles


Related Articles