Paladin AI Looks to Improve Airline Pilot Training Using ML and AWS
While receiving his Ph.D. in Astrophysics, Mikhail Klassen worked with supercomputers running large-scale simulations and performing data analysis, leveraging computing power to explore scientific questions. “Toward the end of graduate school, I had this itch to do something entrepreneurial,” he says. “I really liked this computational approach to solving problems.”
Around the same time, his father, Adolfo Klassen, was leaving a large flight simulator manufacturer, where he had been the CTO. “He believed that the future of pilot training should be data-driven,” explains Mikhail, “and I surmised that machine learning could be applied to this data to enable personalized training.” But this would require a paradigm shift in a conservative industry.
Recognizing a potential opportunity, the father-son duo combined their expertise and experiences to form Paladin AI, a company that uses machine learning to reinvent the pilot training process. “At the end of the day, it’s still a human instructor that certifies the pilot,” says Mikhail. “But there are lots of inefficiencies in that process.”
Instead of relying only on instructors, Paladin AI makes the process more objective by using flight simulator data to find patterns and identify gaps in proficiency. This process allows Paladin AI to differentiate a novice from an expert and build the necessary tools and dashboards, helping pilots address specific weaknesses without wasting time on things they’ve already mastered.
Paladin AI calls this process “adaptive training,” and overall, it makes the pilot training process more efficient, effective, and economical. “There’s so much room in this system to optimize,” says Mikhail. “We have to move away from this one-size-fits-all approach.”
To enable this type of advanced machine learning and scale, the team at Paladin AI turned to AWS and its extensive list of services, including Amazon DynamoDB, Amazon Aurora to manage their SQL databases, and Amazon SageMaker to prototype algorithms. “We can talk to our clients and say, ‘Look, it doesn’t matter whether you connect one of your simulators or 10 of your simulators. The environment we provisioned for you can dynamically scale to handle any amount of data you push to our cloud.'” says Mikhail.
Many of the benefits of using AWS also come from the nature of the aviation industry, which, perhaps more than any other, is remarkably global. “Because we’re in aviation, we’re not limited to any one geography,” explains Mikhail. “We go where there’s demand. We’re not limited—and that’s another great thing about building on AWS.”
According to Mikhail, AWS not only benefits Paladin AI, but it also helps customers—for instance, by cutting down latency. “If we have a European customer, then we’re provisioning in a European data center, and they can get data back in two or three seconds. That’s amazing when you can do that for your customer and know that even as they add more sims, they’re not going to take a performance hit.”
Mikhail also notes the incredible security benefits of using AWS, particularly for customers. “When you’re building on top of AWS, you can take advantage of all of the built-in security and privacy features,” he says. “The fact that this architecture is building on top of AWS’s servers and systems means you know they’re already best in class.” Knowing their customers’ privacy is being protected means greater peace of mind for everyone.
For now, Paladin AI is focused on the commercial airline industry and helping it innovate following the devastating impacts of COVID-19. Still, the company sees the potential to expand in the future, especially more broadly in advanced air mobility and within the defense sector, while pursuing its mission of making flying accessible to more people.
“Getting your private pilot’s license is expensive,” Mikhail says. “It’s a big reason why a lot of people never finish. So, we want to help open the pilot pipeline to people who want to pursue this as a career but don’t necessarily have the financial resources to pay for their training. We want to address this problem of accessibility.”