If you're looking to embark on an AI project, the first step is to recruit the right team. This can be the most challenging part of the process as specialized AI talent is difficult to find. According to the NYTimes, there are fewer than 10,000 qualified people in the world and universities are only graduating about 100 new candidates each year with the requisite skills. Further complicating matters are the myriad of job descriptions, titles, roles, skills, and technologies used in the industry. What does all the terminology mean? And how do they fit into your recruiting strategies for hiring AI talent?
Machine Learning Engineers
At the center of any machine learning project lie the machine learning engineers. With backgrounds and skills in data science, applied research, and heavy-duty coding, they run the operations of a machine learning project and are responsible for managing the infrastructure and data pipelines needed to bring code to production. Explains eBay VP of Engineering Japjit Tulsi, machine learning engineers must be able to "straddle the line between knowing the mathematics and coding the mathematics."
Supporting the machine learning engineers are data scientists who do not typically ship production code, but rather tackle discrete problems using preexisting data to validate models. They have PhDs in data science or statistics, or backgrounds in computer science, math, and physics. According to Greg Benson, Professor of Computer Science and Chief Scientist at AI firm SnapLogic, "data science people are focused on the algorithm and the analysis; they're not operating on the software side." In the process of developing algorithms and analyses, data scientists also perform the critical task of collecting, cleaning, and preparing data correctly which can be the most time-consuming portion of their work. Abhi Jha, Director of Advanced Analytics at McKesson, admits that "the hard work is cleaning data, the model selection is easy."
Research Scientists / Applied Research Scientists
Research scientists often build on promising data leads uncovered by data scientists or experiment with novel approaches, some of which may have originated from academic or industry research facilities. They are more focused on driving scientific discovery and less concerned with pursuing industrial applications of their findings. Tying these two fields together are the applied research scientists. With backgrounds in both data science and computer science, they are invaluable members of any AI team. Explains Tulsi, "they can both pitch in on data science and write code. Finding a good applied research scientist is worth her weight in gold. "
Distributed Systems Engineer
A distributed systems engineer is a support role focused on managing the challenges associated with large data sets and distributed scale issues. Fred Sadaghiani, CTO at machine learning company Sift Science, describes the position as critical. "You need people that understand how to build a distributed system," he says, "they are incredibly important in helping to operate and productionize these systems."
Building a team for a machine learning project is no easy task. With more and more companies turning to AI solutions, the number of experienced scientists and engineers falls far short of the demand. The exact composition of an AI project team varies depending on the nature and timing of the project, but having expertise running from fundamental research and data to heavy-duty coding and production is vital. "There's an old saying", says Tulsi, "that you can't clap with one hand. In this case, you need all three hands clapping together."