Hunter Frields is a graduate student in computer engineering at George Mason University, pursuing a concentration in machine learning and artificial intelligence while working full time in research and development at WRSystems.
He found his path to the field early, driven by curiosity about how technology works beneath the surface. “Starting as a kid, I’ve been trying to build computers and programs,” he said. That curiosity turned into experimentation, with Frields teaching himself enough Java by sixth grade to create simple modifications in Minecraft. The experience offered a glimpse into what he says is most exciting about computing—the ability to build something new and control how it behaves. “I’ve always wanted to build something,” he said. “I like to be able to be in control of the way things work.”
Location played a crucial role in his decision to pursue graduate studies at George Mason. “I picked it because of its strong computer engineering program and proximity to job opportunities in Northern Virginia,” Frields said. “In Reston, for example, you have Microsoft, Google, and companies that are often looking for AI engineers and programmers.”
Frields says the structure of the university’s computer engineering program makes it possible to be a student and work full time. “George Mason is really convenient for having its classes late in the evening, especially for the master’s program.”
The connections between classroom learning and real-world applications are central to his graduate experience. In one course focused on machine learning for embedded systems, Frields and classmates tackled making powerful machine learning models run efficiently on smaller devices. He helped design a system capable of recognizing everyday objects. The project drew on a large-scale dataset and architecture inspired by MobileNet, a neural network designed for efficient performance.
Looking ahead, Frields hopes to push those ideas even further. In a course this semester, he is developing a research project combining machine learning with signal processing—a key part of his current work in industry. His focus is on improving the ionogram inversion process used in over-the-horizon radar systems, which analyze radio signals that bounce off the ionosphere.
When these radar signals return to Earth, they create visual patterns called ionograms. From those patterns, scientists estimate the ionosphere’s electron density, an invisible and critical factor determining how radio waves travel long distances. Today, that process relies on a technique known as multi-quasi-parabolic modeling, which represents the ionosphere using about a dozen carefully tuned parameters.
“I’m wondering if I could train a machine learning algorithm to recognize those signals a little faster,” he said. “I think it would be interesting to try to apply that together.”
His idea is to use machine learning as a front-end assistant. He would train a model to recognize patterns in real backscatter ionograms and produce a fast, high-quality initial estimate of the ionospheric conditions. That estimate could then be refined using the existing physics-based optimization methods that already work well, making the system faster, more robust, and more adaptable to changing conditions.
Frields’ goal is to remain in research and development, solving complex technical problems and finding new ways to make systems faster, smarter, and more capable. “It’s not like a standard programming job. Sometimes the question is just: we don’t know how to solve this problem. Can you try your hand at it?”
That challenge is exactly what keeps him interested. “I love it,” he said.