At the top of LinkedIn's 2017 US Emerging Jobs Report were two occupations in the Machine Learning field: Machine Learning Engineer and Data Scientist. Employment for machine learning engineers grew by 9.8 times between 2012 and 2017 and data scientist jobs increased 6.5 times during the same five year period. If the trend continues, these occupations will have employment outlooks that surpass many others occupations. With a future so bright, could a job in this field be right for you?
What Is Machine Learning?
Machine learning (ML) is just what it sounds like. This technology involves teaching machines to perform specific tasks. Unlike traditional coding that provides instructions that tell computers what to do, ML provides them with data that lets them figure it out on their own, much like a human being or animal would do. Sounds like magic, but it isn't. It involves the interaction of computer scientists and others with related expertise. These IT professionals create programs called algorithms—sets of rules that solve a problem—and then feed them large sets of data that teach them to make predictions based on this information.
Machine learning is a "subset of artificial intelligence that enables computers to perform tasks they haven’t been explicitly programmed to do” (Dickson, Ben. Skills You Need to Land a Machine Learning Job. It Career Finder. January 18, 2017.) It has gotten more complicated, yet more commonplace, over the years. Steven Levy, in an article that speaks to Google's prioritization of machine learning and retraining of the company's engineers, writes, "For many years, machine learning was considered a specialty, limited to an elite few. That era is over, as recent results indicate that machine learning, powered by “neural nets” that emulate the way a biological brain operates, is the true path towards imbuing computers with the powers of humans, and in some cases, super humans” (Levy, Steven. How Google is Remaking Itself As a Machine Learning First Company Wired. June 22, 2016).
How is machine learning used in the "real world?" Most of us come across this technology on a daily basis without giving it much thought. When you use Google or another search engine, the results that come up at the top of the page are the result of machine learning. The predictive text, as well as the sometimes maligned autocorrect feature, on your smart phone's texting app, are also a result of machine learning. Recommended movies and songs on Netflix and Spotify are further examples of how we use this rapidly growing technology while barely noticing it. More recently, Google introduced Smart Reply in Gmail. At the end of a message, it presents a user with three possible replies based on the content. Uber and other companies are currently testing self-driving cars.
Industries Using Machine Learning
The use of machine learning reaches far beyond the tech world. SAS, an analytical software company, reports that many industries have adopted this technology. The financial services industry uses ML to identify investment opportunities, let investors know when to trade, recognize which clients have high-risk profiles, and detect fraud. In health care, algorithms help diagnose illnesses by picking up abnormalities.
Have you ever asked the question, "why is an ad for that product I'm thinking of buying showing up on every web page I visit?" ML allows the marketing and sales industry to analyze consumers based on their buying and search histories. The transportation industry's adaptation of this technology detects potential problems on routes and helps make them more efficient. Thanks to ML, the oil and gas industry can identify new energy sources (Machine Learning: What It Is and Why It Matters. SAS).
How Machine Learning Is Changing the Workplace
Predictions about machines taking over all our jobs have been around for decades, but will ML finally make that a reality? Experts forecast this technology has and will continue to alter the workplace. But as far as taking away all our jobs? Most experts don't think that will happen.
While machine learning can't take the place of human beings in all occupations, it could change many of the job duties associated with them. "Tasks that involve making quick decisions based on data are a good fit for ML programs; not so if the decision depends on long chains of reasoning, diverse background knowledge or common sense” says Byron Spice. Spice is Director of Media Relations at Carnegie Mellon University's School of Computer Science (Spice, Byron. Machine Learning Will Change Jobs. Carnegie Mellon University. December 21, 2017).
In Science Magazine, Erik Brynjolfsson and Tom Mitchell write, "labor demand is more likely to fall for tasks that are close substitutes for capabilities of ML, whereas it is more likely to increase for tasks that are complements for these systems. Each time an ML system crosses the threshold where it becomes more cost-effective than humans on a task, profit-maximizing entrepreneurs and managers will increasingly seek to substitute machines for people. This can have effects throughout the economy, boosting productivity, lowering prices, shifting labor demand, and restructuring industries (Brynjolfsson, Erik and Mitchell, Tom. What Can Machine Learning Do? Workforce Implications. Science. December 22, 2017).
Do You Want a Career in Machine Learning?
Careers in machine learning require expertise in computer science, statistics, and math. Many people come to this field with a background in those fields. Many colleges that offer a major in machine learning take a multi-disciplinary approach with a curriculum that includes, in addition to computer science, electrical and computer engineering, math, and statistics (Top 16 Schools for Machine Learning. AdmissionTable.com).
For those who are already involved in the Information Technology Industry, the transition to an ML job isn't a far leap. You may already have many of the skills you need. Your employer may even help you make this transition. According to Steven Levy's article, "currently there aren’t a lot of people who are experts in ML so companies like Google and Facebook are retraining engineers whose expertise lies in traditional coding."
While many of the skills you developed as an IT professional will transfer to machine learning, it may be a bit challenging. Hopefully, you stayed awake during your college statistics classes because ML relies on a strong grasp of that subject, as well as math. Levy writes that coders have to be willing to give up the total control they have over programming a system.
You are not out of luck if your tech employer isn't providing the ML retraining Google and Facebook are. Colleges and Universities, as well as online learning platforms like Udemy and Coursera, offer classes that teach the basics of machine learning. It is crucial, however, to round out your expertise by taking stats and math classes.
Job Titles and Earnings
The primary job titles you will come across when looking for a job in this field include machine learning engineer and data scientist.
Machine learning engineers "run the operations of a machine learning project and are responsible for managing the infrastructure and data pipelines needed to bring code to production." Data scientists are on the data and analysis side of developing algorithms, rather than the coding side. They also collect, clean, and prepare data (Zhou, Adelyn. "Artificial Intelligence Job Titles: What Is a Machine Learning Engineer?" Forbes. November 27, 2017).
Based on user submissions from people working in these jobs, Glassdoor.com reports that ML engineers and data scientists earn an average base salary of $120,931. Salaries range from a low of $87,000 to a high of $158,000 (Machine Learning Engineer Salaries. Glassdoor.com. March 1, 2018). Although Glassdoor groups these titles, there are some differences between them.
Requirements for the Machine Learning Jobs
ML engineers and data scientists do different jobs, but there is a lot of overlap between them. Job announcements for both positions often have similar requirements. Many employers prefer bachelor's, master's, or doctoral degrees in computer science or engineering, statistics, or mathematics.
To be a machine learning professional, you will need a combination of technical skills—skills learned in school or on the job—and soft skills. Soft skills are one's abilities that they do not learn in the classroom, but instead are born with or acquire through life experience. Again, there is a great deal of overlap between the required skills for ML engineers and data scientists.
Job announcements reveal that those working in ML engineering jobs should be familiar with machine learning frameworks like TensorFlow, Mlib, H20 and Theano. They need a strong background in coding including experience with programming languages such as Java or C/C++ and scripting languages such as Perl or Python. Expertise in statistics and experience using statistical software packages to analyze large sets of data are also among the specifications.
A variety of soft skills will allow you to succeed in this field. Among them are flexibility, adaptability, and perseverance. Developing an algorithm requires a lot of trial and error, and therefore, patience. One must test an algorithm to see if it works and, if not, develop a new one.
Excellent communication skills are essential. Machine learning professionals, who often work on teams, need superior listening, speaking, and interpersonal skills to collaborate with others, and must also present their findings to their colleagues. They should, in addition, be active learners who can incorporate new information into their work. In an industry where innovation is valued, one must be creative to excel.