researcher at zyphra
currently working on novel hardware aware pretraining architectures for low latency inference at scale, diffusion language models, sample efficient context extension, spectral clipping optimizers
-Rishi

Feel free to email me, I am almost always interested in meeting new people.