n8n.io: How Dropsolid enhances digital experiences with workflow automation
The application of artificial intelligence (AI) and machine learning (ML) technology is growing rapidly across disciplines and industries. A challenge is that the demand to apply the technology exceeds the available people with the skills needed to make it useful. There are a number of movements taking place to help address this skills gap, and there are also some trends working against the long-term goal of building a healthy AI workforce that I hope will change. In this article, I will describe both the positive and negative influences affecting the skills gap.
First, some background on my perspective. I’m a Professor of Computer Science at the University of San Francisco and Chief Scientist at SnapLogic, a provider of a cloud-based platform for data integration and automation. At SnapLogic, I started our own machine learning projects about four years ago and have gone through the process of making AI successful in our product for the benefit of our customers. This experience has helped me understand how to cultivate AI talent and make internal changes needed to support AI.
Universities and colleges all over the world are responding to the need for more data scientists. As such, we have seen several new programs at the undergraduate and graduate level focusing on data science, which includes AI and ML. For example, in addition to adding an undergraduate major in data science, the University of San Francisco started one of the first one-year intensive masters programs in data science. Students take a year of courses and participate in a nine-month practicum in which they work on data science projects at a company. While the students may not have a lot of experience initially, they get immersed in practical courses and the application of their knowledge to real business problems. All of these students find jobs immediately after graduation and many of them end up working for their practicum sponsor.
Data science requires a mix of mathematics, statistics, and computer science. I am also seeing an increasing amount of AI and ML make its way into computer science curricula. Beyond offering courses in AI and ML to computer science students, other standard software development courses are beginning to incorporate ML projects because they help motivate students, and AI and ML will increasingly become a part of how many software engineers create software. So, in addition to specifically trained data scientists, emerging computer science graduates will have more and more exposure to AI and ML.
One way to find data science talent is to look to local colleges and universities that offer data science or computer science programs with exposure to machine learning. These programs will often have a practicum or project course in which students work with industry on the application of machine learning. Becoming a sponsor can lead to an opportunity to hire AI talent. This academia/industry collaboration has worked well at SnapLogic, with several University of San Francisco graduates now working on SnapLogic’s engineering team.
Of course, there is also a plethora of online courses and tutorials focused on AI and ML. Existing analysts, statisticians, and computer programmers are actively getting up to speed and, in many cases, companies are encouraging and paying for existing employees to retrain through these online resources. Very large companies, like Google and Facebook, have created their own internal training programs.
Unfortunately, in the race to innovate in AI technology, the large well-funded companies are also aggressively hiring many top machine learning PhD students who may otherwise choose an academic career. While it is normal for PhD students to opt for industry, it appears we may be seeing an imbalance towards industry due to extremely high salaries and the opportunity to work on challenging, real-world problems. However, we really need many of these PhD students to continue an academic track so that they can help train future generations of AI experts. In this regard, one solution I see is for the established AI-driven companies to encourage their AI experts to teach ML and AI courses at local universities in addition to sponsoring student projects.
Finally, another way to tackle the AI skills gap is to invest in software that can assist in the creation of machine learning models that can be applied to improving your customer experience or business processes. Both code and no-code options are still in their early stages and they all show promise. For example, all of the big cloud service providers have APIs and web-based tools for assisting in the creation of machine learning models. Other companies, like SnapLogic, provide a visual drag-and-drop approach to building a machine learning workflow for model training and deployment. All of these options will make it easier for businesses to take advantage of AI technology more rapidly as opposed to doing everything from scratch.