Last year saw a record number of investments in AI with $5 billion in funding, and the number continues to grow. Investors’ activity indicates the significance of this kind of technology for society – self-driving cars of the future, DNA genome analysis, climate change, cancer predictions and many other fields. Besides the important role AI plays in these fields, AI as a technology is simply more efficient than traditional technology. In our company, we needed two weeks of training and almost no human involvement to introduce new AI filters for images. Before, it would have required the work of two to three engineers and at least two months of development. If we look at another example, relatively young AI-driven cybersecurity companies like Cylance or Lookout compete heavily with veterans like McAfee because AI is an integral part of their products. Even tech giants, in some cases, concede to startups in the field of AI. As a recent example, Russian startup NTechLab beat Google in the“MegaFace” facial recognition competition.
Such productivity in the practical implementation of AI will continue to fuel the high demand for data scientists, machine learning engineers, ML researchers and all other professions related to the field, which will effectively replace computer science altogether. Moreover, companies that are operating in different verticals – such as image recognition, voice recognition, medicine or cybersecurity – are already faced with the challenge of acquiring a workforce with the right set of skills and knowledge.
A traditional computer science engineer is not able to solve those tasks, so the demand for a new skill set is growing, especially in regard to data scientists, for whom the demand is projected to exceed supply by more than 50% by 2018. This is probably a good indication as to why Harvard Business Review declared data scientist to be the “sexiest” job of the 21st century back in 2012. A data scientist’s biggest skill is the ability to formulate a question from data and understand the context the data is gathered from. Computer science work is a logical process, but most data science work is an exploratory process, which is why, because of the boost in AI technology, this scope of work is in demand. But universities and other educational organizations are simply not able to keep up with such rapid changes.
To stay competitive, companies need these specialists now and cannot wait five years for universities to produce graduates from new courses. The fastest route is to retrain graduates in math and physics, the specialties that are strong in statistics. My company is already running such a training program and building a new sort of data science/ML school in Armenia — a country that is already strong in the sciences. The program was launched at the end of 2015 and has so far graduated 400 students, and we were able to hire 50 of them.
Such programs have opened the door to a global market that urgently needs professionals especially since, one in three data scientists in the U.S. are foreigners. This approach has proved to be helpful since students are taught real problems using real data. They are prepared for a wide scope of tasks and challenges.
Co-founder and CEO of PicsArt, a leading social photo editing startup with more than 90 million monthly active users, Hovhannes Avoyan.
The article initially appeared in Forbes.
