Maximizing Human Performance with AI

When people ask AI Program Manager Chris Shumeyko how he got to where he is today, he often describes himself as an “accidental AI engineer.”

While his education has focused on mechanical engineering, other experiences—ski racing, Crossfit, leading a sporting association, and more—have contributed to his career path and to studying human performance, predictive maintenance, and AI adoption at Booz Allen. It’s all about using data to solve problems, Chris says.

Read on to see how all the pieces come together.

“I’ve always been interested in numbers and analytics,” Chris says, describing his bachelor’s and PhD in mechanical engineering.

For his PhD in computational material science, he explored how atoms and molecules move and “affect the world around us. I was looking at very, very small things and their application to the real world and big objects.”

Chris describes his early work as “AI adjacent,” dealing with numbers and analytics. “At the time, 2010 to 2015, AI had only started to enter the mainstream,” he notes.

“AI isn’t magic,” Chris stresses, “it’s just math. We’ve been doing it for years.”

After completing his Ph.D., Chris started a postdoc at the Army Research Laboratory. There, someone approached him about an opportunity with the DOD’s newly established Joint AI Center (now known as the Chief Digital and AI Office).

At first, Chris was hesitant. “I didn’t know a thing about AI,” he says. But he soon saw that the overlap between his mechanical engineering background and AI involves fixing things. “So, I decided to try it,” he continues.

The role proved to be a good fit.

“Working as a tech advisor in the predictive maintenance space, I was a kind of interpreter, explaining AI to generals in a non-technical way.”

A client project enabled Chris to work on maintenance and logistics, helping AI adopters understand their specific problems. This laid the groundwork for his current approach to demystifying AI.

“It’s where I discovered that I could take my technical background and communicate with non-technical audiences,” he says. 

Meanwhile, Booz Allen was looking for communicators like Chris, who can understand and translate people’s problems and use technology to find innovative solutions.

“I had one conversation with Booz Allen’s CTO (Chief Technology Office) team,” he recalls, “and they just understood what I was trying to do.”

From his home in Pittsburgh, Chris joined Booz Allen’s predictive maintenance team, where he leads the development of AI-enabled supply chain and predictive maintenance solutions, focusing on scalability and transparency. “We want to leverage tech like AI to make lives easier. We’re exploring how we use data to better predict when things may go wrong, to avoid critical situations,” he explains.  

Many federal agencies know that AI can transform their work. But they don’t know where to start. That’s where Chris and his team come in. They use data to identify pain points and solve problems.

One of his current projects is working on developing a product suite focused on enabling the AI journey.

“There’s data and information everywhere, but it’s not easy to find,” he says. With this product suite, “we’re going after the low-hanging fruit, taking away the mundane, day-to-day tasks and making people’s jobs easier, so they can start to do the fun stuff.”

Chris is also working to ensure that the human element doesn’t fall by the wayside. “The human aspect brings so many unique perspectives to solve problems,” he says. “If everybody has the same perspective, they’re going to solve the problem the same way.

“Our team is made up of diverse individuals,” he notes. “We have people with finance degrees, PhDs in psychology, engineers, data scientists. It’s not just 50 computer scientists building AI tools.

“I think of how Booz Allen is different in the AI space; it’s that we understand the people we work with and who we’re trying to help. We define AI as human—anybody in the world can develop an AI algorithm or a product to try to solve a problem.  But if they don’t understand the problem they’re trying to solve right, and the people they’re trying to work with, that tool is not going to be useful.”

During a recent talk at Mississippi State University, Chris described the role of data in improving human performance. Data can help answer questions about training most efficiently and ensure top performance.

To illustrate his point, Chris reflects on his time as a ski coach, working with college athletes. “As a coach, a lot of my time was dedicated to helping new athletes,” he says. “New students don’t understand their body or how to move properly, so that’s what I’m going to focus on. But doing that means that I’m ignoring 75% of the team.”

With AI, Chris can coach by reviewing AI insights and discussing them with the team’s more experienced skiers. In this case, AI isn’t replacing the coach’s role but providing visibility into the performance of the larger group.

But Chris cautions, “we must be careful to always keep the human in the loop to prevent us from truly just trusting an algorithm or a tool.”

“If we solely rely on AI, that takes away the compassion that a coach or a mentor could bring to a problem,” he says. “We have to continue to ground ourselves in that subject matter expertise to make sure the solutions we develop to make us more efficient data-driven societies don’t get out of hand.”

Booz Allen combined AI, human performance technology, and expert coaching at the 2024 DOD Warrior Games Challenge, an annual event where wounded, ill, or injured veterans and active-duty athletes representing all military branches come together to compete. Booz Allen coaches analyzed personalized physical and cognitive performance metrics to help athletes using wearable devices optimize their training. The same integrated approach to data analytics is used to support the DOD by assessing the readiness, well-being, and performance of warfighters in the field.