Tirth Patel

I'm Tirth, an ML compiler engineer at NVIDIA. I work on PyTorch and the compilers, runtimes, and hardware paths underneath it.

Before NVIDIA, I spent a year and a half at Tesla working on Dojo and Autopilot: compilers, performance work, and the tooling that helped us scale to thousands of GPUs.

I've also spent a lot of time in open source: Keras as a student researcher at Google, SciPy (I authored scipy.stats.sampling), PyMC, and Google Summer of Code.

I care about systems that are fast, legible, and pleasant to use.

I'm always happy to talk about compilers, GPU performance, or open source. Email is the fastest way to reach me.

• Email
• GitHub
• LinkedIn