I work on machine translation and language modeling as a Senior Research Scientist with Bloomberg Labs.

Previously, I worked as a Postdoctoral Scholar at Stanford, as a Research Associate at the University of Edinburgh with Philipp Koehn, and as a PhD student at Carnegie Mellon advised by Alon Lavie.

My recent papers cover three areas:
Hypergraph search
 is a faster and more accurate alternative to cube pruning.
KenLM
 estimates and queries language models using less time and memory. It is behind three entries to the 2013 WMT evaluation, all of which won first place in their respective language pairs.
MEMT
 won best system combination in 8 of 10 tracks of the 2011 WMT evaluation.

I like travel, C++, math homework, and networking.

This personal website does not represent the views of my employer.