Research Scientist · Google

Ashwini Pokle

My research aims to make generative models more capable and efficient by developing algorithms and architectures that generate higher quality images with lower computational cost. I work across sampling, training, and model design to improve the fidelity, controllability, and alignment of generated media.

I received my PhD from the Machine Learning Department at Carnegie Mellon University, advised by Prof. Zico Kolter. My thesis explored efficient generative inference by combining Deep Equilibrium Models and Diffusion Models. Prior to that, I obtained my MS in Computer Science at Stanford University. I also spent time at Stanford Vision and Learning Lab, where I worked on behavioral and social robot navigation under supervision of Prof. Silvio Savarese. I earned my BE (Hons.) in Computer Science from BITS Pilani, India. I have also interned at FAIR (Meta AI) and Bosch AI.