I study computer engineering at Southern Miss.
Lately: efficient inference, ARM kernels,
and keeping a Raspberry Pi from overheating.
01 / RESEARCH & EXPERIMENTS
Papers, code & ongoing work.
Keeping computer vision running without a fan. An empirically calibrated DVFS scheduler for the Raspberry Pi 5 that balances thermal limits, throughput, and energy use.
A neural-network inference engine built to understand the hardware underneath. Hand-written ARM NEON kernels, a zero-allocation hot path, and an optimization ladder measured on the Cortex-A76.
8.45× faster than the naïve baseline, with the stable zero-allocation NEON path.
Reconstructing the Ξc+ baryon across three decay channels in Belle II simulations. A collaborative pipeline from Monte Carlo generation to machine-learning event selection.
8.5M+ simulated events. 3.3 GB of openly available data.
02 / ABOUT
I’m Aayush, a Computer Engineering student with a Mathematics minor at the University of Southern Mississippi. My interests sit where computer architecture, memory systems, and efficient AI meet.
I like figuring out what’s slowing things down and making better use of the hardware that’s already there. Mostly by building things, measuring them, and trying again.
Beyond the lab, I’m a two-time national mathematics olympiad finalist and represented Nepal at the 2025 IOMHSS in Turkmenistan. I also founded VIDYA, helping students across Nepal discover academic opportunities.
Outside of work, I like playing football. I support Real Madrid.
03 / CONTACT
For research, projects, or a quick hello.