Part of:How data science jobs will look in the future
Data scientists vs. machine learning engineers
The positions of data scientist and machine learning engineer are in high demand and are important for enterprises that want to make use of their data and use AI.
Any job that works with data is a high-profile position these days, with huge demand and commensurately top salaries for such roles.
Data scientists and machine learning engineers are among those top occupations, fueled by enterprises' need to make data-driven decisions.
"The prospect for both jobs is very rosy," said Lian Jye Su, who, as a principal analyst at ABI Research, is responsible for orchestrating research relating to robotics, artificial intelligence and machine learning.
Still, the positions require different skills and take on separate responsibilities in the enterprise.
The data scientist role
At the most basic level, a data scientist analyzes structured and unstructured data to generate information and gain insights that organizational leaders use to answer questions, solve problems or predict outcomes.
Although the role has emerged as one of the hottest jobs in the 21st century, it draws on a longstanding practice among business executives and organizational leaders to use data to make decisions.
Of course, organizations today need to manage and analyze huge amounts of data. Data volumes saw explosive growth starting around a decade ago and that growth has continued to accelerate.
That volume -- along with technologies used to capture, store, manage and analyze the data -- has given rise to numerous data-related positions, including database administrator, data engineer and data scientist.
The roles aren't always clearly defined, said Scott McClellan, senior director with the data science product group at Nvidia, a manufacturer of graphics processing chips.
"The terminology in this space is quite imprecise," he said.
In fact, McClellan said the roles and responsibilities assigned to these titles often vary from one organization to the next.
"They're all subject to a little case-by-case variation and some overlapping because this space is under constant innovation and evolution," he said.
"As the name suggests, data scientists are responsible for assisting companies to adopt data-driven decision-making via discovery of valuable insights in data," Su said.
A data scientist uses complex algorithms, manipulates data and leverages a host of technologies that require specialized math, technology and computer skills, and business acumen to generate insights.
"They will lead the discovery process, work through extracting, cleaning and loading the data before conducting investigations and analysis of the data by applying statistical models or algorithms to the data set," added Jim Johnson, senior vice president for the technology division of Robert Half Technology, a human resources consulting firm.
A data scientist must be capable of the following:
acquiring and collecting data;
processing and cleaning data before storage;
setting up the right data streaming and collection architecture;
initializing data investigation and exploratory data analysis through data science techniques such as statistical modeling and machine learning; and
present findings and insights to key stakeholders and adjust based on feedback.
According to the Robert Half Technology 2021 Technology Salary Guide, annual pay for data scientists ranges from $109,000 to $186,000. Meanwhile, the U.S. Bureau of Labor Statistics puts the salary range for data scientists and other mathematical science occupations at $53,000 to $158,000.
The U.S. Bureau of Labor Statistics also identified data scientists among the 12 fastest-growing occupations for the upcoming decade. Computer and information research scientists (a grouping that includes machine learning professionals) are expected to grow 15% over the decade, significantly faster than the average professional category.
The machine learning engineer career
Machine learning engineers also work with data, but in different ways than data scientists. They rely more heavily on programming skills than other data-related positions do.
The prospect for both jobs is very rosy.
Lian Jye SuPrincipal analyst, ABI Research
At a basic level, a machine learning engineer creates software programs and algorithms that allow computers to perform actions without being directed or told what specific steps to take.
"A machine learning engineer is often involved in the same projects as a data scientist, but comes at it from a different perspective," Johnson explained. "While a data scientist will analyze and research data, an engineer will build the software or platforms that will continue to enable the functionality in production. They often sit between software engineers and data scientists."
To do that work, a machine learning engineer needs to have the following:
a strong background in data science;
software engineering skills;
experience with the programming languages Python, Java, R, JavaScript and C/C++;
the ability to design and develop machine learning systems based on open source machine learning frameworks, SDKs and libraries;
the capability to engage in continuous testing, validation and versioning; and
an understanding of microservices architecture and container management as well as open source communities and open source technologies such as Linux and the container-orchestration system Kubernetes.
How much a machine learning engineer earns
Like data scientists, machine learning engineers are in high demand. According to a survey by Robert Half Technology, 30% of U.S. managers said their company already uses AI and machine learning and 53% expect to adopt these tools within the next three to five years.
Since the position is so new, Robert Half Technology points to other similar positions when estimating the salary range for machine learning engineers, with the expected annual pay ranging from low to mid $100,000s. Indeed.com put the average annual salary at about $151,000.
Overlapping work, responsibilities
Data scientists and machine learning engineers collaborate in most organizations; some might see an overlap in their roles and responsibilities as well.
The two share work, particularly in the areas of data preparation and management, as well as in the model selection phase, and see their positions closely connected throughout projects.
"In a typical machine learning model lifecycle, data scientists will play a heavy role in the beginning, before handing over most of the operational tasks to machine learning engineers during operation," Su explained. "The data scientist will return to the picture once the machine learning model has generated insights that require analysis and breakdown."
