About

My name is Anand. I like to work at the intersection of mathematics, physics, and computing. Lately my interest in compilers and low-level performance has taken me to Modular, where I work on end-to-end inference performance. Kernels are one part of it, written in Mojo and tuned across accelerators. The rest is everything else that sits between a fast kernel and a fast model. I poke at ML compilers in my free time.

Before that I helped build a GPU-based fluid dynamics solver at Ansys, and improved physics-based models with data at Geminus.

I studied at the University of Michigan (PhD, aerospace engineering) and IIT Bombay (master's and bachelor's, aerospace engineering), alongside some of the brightest people in the field.