911������

When Cameron Jozwik ‘27 presented his paper at the IEEE Conference on Secure and Trustworthy Cyberinfrastructure for Internet of Things and Microelectronics (SaTC 2026), he presented alongside researchers working on security challenges in next-generation connected systems. For the 911������ Computer Science student, the moment marked a significant milestone in his academic progression.

Jozwik presented his paper, Passive On-Device BLE Interference Detection Using TinyML, as part of the conference’s “AI for Smart Environments, Manufacturing, and Autonomous Systems” track. His research tackles a common but often overlooked problem: unreliable Bluetooth performance in crowded or signal-heavy environments.

At its core, the project introduces a lightweight embedded system that enables devices to detect interference in near real time without relying on external infrastructure or added hardware.
For Jozwik, this represented a clear turning point:

“Getting a paper accepted to SaTC 2026 shows that my work has moved beyond a typical class project into something recognized by the research community. At this point in my time at Kettering, it is a shift from just learning concepts to actually contributing to the field.”

Solving a Hidden Problem in Everyday Technology

Bluetooth Low Energy (BLE) operates in a shared wireless spectrum, which means interference is not just possible; it’s constant. Devices compete with Wi-Fi, other Bluetooth signals, and environmental noise, often without awareness of what’s actually happening.

The practical impact is immediate. From consumer electronics to industrial Internet of Things (IoT) systems, the ability to detect interference at the device level opens the door to more resilient, responsive technology.

And it connects directly to real-world experiences people already recognize. Anyone who has experienced earbuds cutting out, devices disconnecting, or smart technology lagging in a crowded space has seen the problem firsthand.

“Cameron’s work goes beyond a classroom project and delivers a system that runs directly on embedded hardware under real constraints,” said Rui Zhu, Assistant Professor of Computer Science and co-author of the paper. “Instead of just demonstrating an idea, he built a practical, on-device solution for detecting BLE interference. This focus on real-world deployment and scalability makes the work meaningful to the broader research community.”

Building for Real Constraints

What makes the work stand out is not just the problem it tackles, but how it does so. The system is designed to run on an ESP32 using TensorFlow Lite Micro—a combination that demands efficiency at every level.

That constraint shaped the entire development process. Jozwik noted the hands-on demands of the work:

“I had to collect real wireless data, identify which patterns actually matter, build and test a model that could run on limited hardware, and address hardware issues on the ESP32. I also had to evaluate results and refine the approach when things didn’t work as expected.”

The result is a system capable of near real-time detection with minimal resource use—an important step toward making embedded security both practical and scalable.

From Foundation to Capability

Jozwik’s ability to take on a project at this level reflects the blend of rigorous coursework, applied learning, and faculty mentorship that defines the Kettering experience.

He pointed to his academic and mentoring experiences as key preparation, noting that his coursework supported programming while mentoring helped him “train a Machine Learning model.”

Over the past year, that foundation has translated into measurable technical growth. He emphasized that he has learned to use security tools and analyze BLE and Wi-Fi packets.

That evolution, from learning systems to analyzing and improving them, mirrors the demands of embedded security and connected systems.

The work presented at SaTC 2026 is not an endpoint. It signals where his career is headed. Jozwik noted that the experience “helps me understand Embedded Security and practices used within it.”

In a field where reliability, security, and real-time performance are increasingly intertwined, that direction matters. Systems are no longer judged solely on what they can do under ideal conditions, but on how they respond when conditions are unpredictable.

His research sits squarely in that space: focused, practical, and built to operate in the real world.

For Jozwik, SaTC 2026 was more than recognition. It was evidence that undergraduate students can do more than prepare for the future. They can help build it now.