For as long as the introduction of machine learning has been discussed, the inevitable conclusion has always been that people will lose their jobs or their roles will change. In fact, it was a major part of the agenda at Davos this year. And the standard party line has always been that their roles will become more “strategic.” A consequence that has been heralded as an inherent positive, although rarely with any explanation as to what that role will look like.
It leaves the IT engineer feeling frustrated and uncertain about the future.
It’s time to set out exactly what the new strategic roles could look like. Satya Nadella, CEO at Microsoft, expressed a similar urgency at Davos when he said, “Clearly the thing that is top of mind for all of us given the political cycle, is if surplus is going to get created [by AI], I think we’ve got to talk about how the surplus is distributed.”
Obviously, the formulaic role of “fixing” the IT support ticket is going to disappear. But the experience-driven role of MSP business management and consultancy will naturally survive. So, where does this leave the IT engineer in his or her early twenties who is unlikely to have the experience to shift towards being a strategic consultant, outsourced CIO, or trusted advisor?
What will the IT engineer of the imminent future (and in some cases, the present) look like?
Starting with data
Data analytics is the foundation on which machine learning is being built, and is going to be where the role of the IT engineer evolves. For IT engineers, it is going to be crucial to differentiate between the distinct types of data.
Essentially, there are two types of data analytics: predictive analytics and prescriptive analytics. Predictive analytics is key to machine learning, as it can utilise both historical and recent data sets to gauge the probability of outcomes, as well as highlight any uncertainties associated with the methodologies used in forecasting. This enables a human or a machine to make informed decisions to enable (or prevent) a future outcome.
Prescriptive analytics then codifies these suggested actions in the form or rules, policies and/or constraints.
The role of IT engineers will therefore be to sit in the middle of these data types, and determine how data goes from predictive to prescriptive analytics. By using tribal knowledge, engineers will be responsible for determining whether an action is in fact automated.
The three Vs
IT engineers of the future will also need to understand the data on which machine learning is based—its respective uses, limitations, and how to manage the different types of data analysis.
To oversee the mechanics of machine learning, the same three Vs that apply to managing typical big data projects can be readily applied, albeit with some subtle edits.
The first V is volume. How much data do you need to start a machine learning project? While machine learning is still in its infancy, the technical limitations have not yet been fully explored, which means gray areas and even outright confusion still exists. IT engineers therefore need to understand how to hold all this data, be sensibly selective, as well as understand how to query against it.
The second V is the variety of data. Typically, in a business, there are so many different types of data—video, audio and textual search—that are just as valuable as standard numeric data, and these sources need to be included in any machine learning-based analysis. The IT engineer’s new role will include working out how “non-traditional” data sources can be included in analysis to avoid decisions being made based on partial pictures. For example, if a sales department were to record its calls, think how valuable it would be to convert the audio into text, and have algorithms discovering what products are most often sold together, or to see what makes the difference between a successful and unsuccessful sales call.
The third and final V is of course the velocity of data, which deals with the speed of data creation and processing. Much of machine learning’s practical application is reliant on accessing and processing the most up to date information and delivering actionable conclusions while the original data is still current. If an IT engineer is tasked with overseeing the practical use of machine learning-based tools, then maintaining speed and quickness of response and reaction will be a critical measure of success.
The power of search
However, the three Vs are not sufficient. To make full use of machine learning and the underlying data, and therefore capitalise on its career-boosting potential, there is one additional factor that is often forgotten—search.
Search allows a person to satisfy their curiosity and inventiveness quickly. The “what if?” question that suddenly occurs to an engineer, and is a hypothesis that sits outside the bounds of the machine learning model, can yield valuable insights. This means that if IT engineers can begin to unleash their creativity and lateral thinking—not a trait that is often associated with the IT engineer of the recent past—they can use the data and predictive power of machine learning to determine how small, seemingly unrelated changes to IT infrastructure can deliver important benefits.
The human touch
While machine learning will be proactive and help mitigate risks, it doesn’t mean that the role of the IT engineer isn’t still needed. Machine learning is nothing more than a tool. No matter what “thought processes” are being assigned, a person is still needed to ask the question, interrogate the answer or apply it to the business.
It will be the role of the IT engineer to learn about the capabilities and potential uses of the data sets, the sorts of conclusions that can be drawn from data using machine learning models and most importantly, to adapt his or her own capabilities to take maximum advantage of the tool. The danger is that otherwise, the IT engineer becomes simply a person that inputs other peoples’ queries and theories into the model, which is a role that will have a short life expectancy.
Joe Kim, Senior Vice President and Global Chief Technology Officer, SolarWinds
Image source: Shutterstock/Sarah Holmlund