Our customers - mostly Fortune 1000 organizations - are in some stage of a machine learning project. And it's early days, even in these very large and well-funded organizations. There aren't a lot of ML models in actual production today.
This will change. As Gartner and others have reported, AI is a top 3 priority for most enterprises. In order to progress as Gartner presumes enterprises will, they will need to sync their investments up with their priorities. AI expertise is one such investment.
Organizations in these early stages are trying to close their most obvious skills gaps by hiring individuals with advanced AI and ML degrees. The problem is that these people are in very high demand and short supply.
There are other roles that, in our experience, organizations should be treating with more urgency. AI project management is a particularly pressing need. Unfortunately, professionals with experience driving enterprise ML project to successful deployment and production are also in short supply.
It's not unusual for enterprise ML project teams to navigate a project's proof of concept stage with a core group of data scientists. But we've found that when teams exit that stage, when they win approval to build out a production ML system, they begin a process of determining which skills they want to develop as part of their in-house AI core competencies.
Some organizations will decide that they want to do it all internally. In a recent research report Gartner recommended that some aspects of ML system development be routinely outsourced. Topping the list of activities that organizations told Gartner they outsource is "data collection and preparation." Fully 24% of Gartner's respondents indicated that their practice is to find partners that have the expertise to produce training data with high quality at scale.
It’s early days for ML organizations and expertise is still quite low for newly created roles for this specialized discipline. An ML project can be successful without bringing all of this expertise in-house. Determine which core competencies are most important to your ML organization and hire and partner accordingly.