It is not uncommon for a new profession to emerge almost overnight.
And it is not uncommon for many people who are trying to get into this new field and have not been formally trained in regular higher education institutions.
In the area of machine learning and related data science, the path to career progression is quite different from that of most other occupations in the past. It represents not only one of the most promising jobs in the next few years, but also a model of how people entering the workplace today can adapt to future job needs.
Machine learning refers to a technology that combines statistics with computer science and has revolutionized the field of artificial intelligence. Machine learning relies on a new class of learning algorithms that evolve over time, as well as the availability of large amounts of data to train the system.
Many organizations have already business development and invested in the IT infrastructure needed to bring “big data,” typically breaking down internal fragmentation and storing data more centrally. The need to better capitalize on these assets has led to a bout of demand for professionals that goes well beyond the talent trained by traditional computer science programs.
Machine learning is not open to computer scientists. As a discipline centered on data collection, collation, and analysis, machine learning spans many specialized fields and mathematics, statistics and programming are all part of it.
Many non-expert executives also find it hard enough to know one or two in this area because they need to interact with the first-in-class machine learning specialists.
No single job description covers this emerging field completely. Many traditionally known as “data analysts” are eager to win the title of “data scientist,” said Anthony Goldbloom, founder and chief executive officer of Kaggle. His company has an informal network of data experts around the world and was acquired by Google. He added that data scientists, in turn, are eager to become machine learning experts.
The demand for these skills has exploded, at a rate that has made it impossible for those trained in traditional academic curricula to meet their needs. According to Stack Overflow, a community of online developers, data scientists, machine learning experts, and developers with a statistical or mathematical background occupy three of the top four positions in the software industry at the highest pay rates. Stack Overflow’s annual survey of this area is one of the most extensively covered in its survey.
People are entering this area through some unconventional means. A recent survey of 16,000 Kaggle users found that only 30% of respondents had studied machine learning or data science courses in formal university education.
In contrast, 66% of respondents said they are self-taught. Just over half of the respondents said they learned the new discipline through online courses.
People attracted to machine learning come from many different professions. Goldblum cited majors in physics, computer science, classical statistics, bioinformatics and chemical engineering. This makes machine learning the first new discipline to embody the importance of lifelong learning. He said: In the future, it will not be possible to learn new skills throughout one’s entire career to accommodate new opportunities such as these.
The speed with which people in other fields adapt to machine learning also reflects the nature of this discipline.
As a pioneer of deep learning technology, Stanford University professor Andrew Ng said it’s actually easier for nonprofessors to get into the field as the field grows.
“What amazed me was that getting into the world of artificial intelligence was so easy as a business development with the rise of deep learning, the algorithms we used became simpler and easier as we relied on the data,” he said. “A few weeks later, you will be able to read leading research papers and cutting-edge ideas in the field.”
Ng is also the founder of Coursera, a large-scale MOOCs platform. The AI course he taught himself was the first course to attract a large audience on the Internet. However, in the final study of the people, the last few really completed. Following MOOCs, there have been some more structured and targeted subdivisions of online classes. “The speed of knowledge has been greatly accelerated,” he said.
The new discipline is so fast-shaped that it poses challenges for companies trying to capitalize on these skills. How do they design jobs that take full advantage of these new data experts? How to meet the expectations of these new employees and develop a career path for them toward senior management? The lack of technical capacity of senior management has begun to be a problem for these companies.
The bad news is that, at least so far, many companies seem to have failed. According to Kaggle, most people working in the field say they spend an hour or two a week finding new jobs, says Goldbloom.
This is confirmed by the findings of Stack Overflow’s 64,000 developers. Of the developers who said they were looking for a job, machine learning professionals topped the list, accounting for 14.3%. Followed by the data scientists, accounting for 13.2%.
Goldblum said people who work in this area have a strong sense of frustration. Low data quality is one of the main reasons: employers can not provide them with the key raw materials necessary to get the results.
Others complain that the question of their answers is not clear enough in itself. Companies may smell the opportunity through their business development, however, and they often lack a sufficient understanding of how to fully leverage their data assets. It also highlights the lack of technical knowledge of non-specialist managers who work with data scientists and machine learning specialists.
In addition, the forerunner of any emerging career will be frustrated. Goldblum said people complained about finding no other talent to work with.
Companies that grow on the Internet, collect large amounts of user behavior data as part of their business developments, and continually improve their services using technologies such as A/B testing are the natural attributes of these people. To compete for some key talent for the future, other companies need these technologies as their own core business.
Which countries will become the first centers of machine learning?
Kaggle runs a social network of data scientists and machine learning specialists. A quarter of its user base comes from the United States. Users from India ranked second, accounting for 16%.
However, the global distribution is not complete: for example, the data do not reflect the current status of such talent in China because China’s Great Firewall prevented citizens from joining the communities. However, according to the data in September, the monthly user base of the site reached 165,000, which reflected the early development of this emerging subject from one aspect.
Of these users, 27% work in technology and Internet companies, the largest group. 15% work in the academic field and 14% in the financial and insurance fields.
Kaggle helps businesses connect with people in this emerging field by holding open data science competitions and giving bonuses to the best performers in the competition. An interesting finding is that only 2.5% of Kaggle’s user base comes from Russia, but 9 of the 94 “top minds” – the site’s highest-rated users are from Russia.