Navneet

Hey! I am a third-year PhD student in Computer Science at Georgia Tech. My research sits at the intersection of machine learning systems and computer architecture, with a focus on algorithm–system co-design for accelerating LLM and diffusion-model inference on GPUs. I am currently working with Prof. Celine Lin on hierarchical-sparsity techniques and custom CUDA kernels to speed up block-diffusion model inference.

Along the way I have worked on GPU kernel scheduling at AMD, ML-based performance modeling for distributed-training simulators with Prof. Tushar Krishna, secure cache design with Prof. Moinuddin K. Qureshi, and side-channel reverse-engineering of ARM CPUs with Prof. Daniel Genkin. I completed my undergraduate in Electrical Engineering at the Indian Institute of Technology, Bombay, where my research on storage-efficient, secure last-level caches led to the Maya Cache (ISCA ‘24).

Besides research, I indulge myself in various sports activities. Table tennis, football, and badminton are among my favorites along with inline skating.

If you are looking for my CV, click here