I'll be at SIAM CSE25 next Tuesday talking about optimizing CUDA kernels for chemistry foundation models - stop by MS82 and the rest of @giulsguidi.bsky.social's minisymposium! meetings.siam.org/sess/dsp_pro...

I'm happy to share our new preprint on equivariant graph neural networks! These machine learning models are popular in computational chemistry. They rotate their output predictions when the input coordinates rotate, but are computationally expensive (a problem we tackle): arxiv.org/abs/2501.13986

If you want an academic website and really like Python, you should try my fork of the al-folio template! github.com/vbharadwaj-b... It's easy to set up, hackable, and based on Python's Pelican package (the original uses Jekyll / Ruby). Check out my personal page (see bio) for a demo :)