I’m Supavit, an engineer in Seattle working across software and hardware. My experience spans analog signal conditioning, embedded firmware, and robotics, alongside backend services, GPU inference, and infrastructure.
I ran an eight-GPU inference setup and tested how to divide it across four model configurations, using memory and throughput measurements to choose the layout.
I started at Gradient working on automated sorting machines, then moved into backend and infrastructure work.
The robotics work involved motion control, cameras, scanners, and software for machine operators. More recently, I’ve worked on databases, search, model serving, and deployment infrastructure.
In an earlier sensing-rover project, I designed the amplifier and voltage-offset circuit between an ultrasonic microphone and an STM32 ADC, then processed the sampled signal with FFT analysis.