I am an ML Engineer at Myrtle.ai, where I develop training pipelines and train automatic speech recognition (ASR) models. Prior to this, I was an ML research assistant at the CBL, Cambridge University. (See my CV).

I am broadly interested in scaling ML models to massive contexts and datasets through algorithmic improvements. Recent work includes: (1) efficient Bayesian inference algorithms that scale to graphs with over 1M nodes; and (2) reducing the HBM footprint of ASR models by 100× via kernel fusion. I am passionate about collaborative research and always interested in new opportunities to work on challenging problems.