My first bridge between theory and engineering was a voice-controlled windshield wiper, a project that tested whether mathematical logic could drive real hardware. During this engineering project supervised by Professor Ning Yu, I faced the challenge of modeling the nonlinear transition of a robotic motor. As we discussed control-system behavior, I realized the curve we needed resembled a sigmoid function—long before I knew it was widely used in deep learning. I derived it independently, then experimented with a modified version that better fit the motor’s response. I learned C++ and used it to program the Arduino. I also wired and connected the Breadboard, Arduino, and Servo motor. Watching a mathematical idea reshape a physical system felt like discovering a new language.