|
Greg HannemanMachine translation and natural language processing |
I'm an independent MT/NLP contractor based in Pittsburgh, Pennsylvania, USA. My recent work contains a number of threads:
HL4AI students: Please note that all course materials are on the CMU Canvas site. My office hours are Fridays 10:30 to 11:30 a.m. in Wean 3002. You can also reach me via e-mail.
I arrived at Carnegie Mellon in the fall of 2005, when I started my graduate research as a master's student in the Language Technologies Institute working in multi-engine machine translation (system combination). As a Ph.D. student, I focused on MT methods that incorporate lingustic knowledge into the model, in the form of probabilistic grammars. My thesis project developed automatic ways of optimizing the part-of-speech and constituent labels used in synchronous context-free grammars in order to improve MT performance. Throughout my time at CMU, my academic advisor was Alon Lavie.
During the fall of 2009, I completed an internship exploring word alignment and phrase extraction methods at the Xerox Research Centre Europe (now Naver Labs Europe) near Grenoble, France.
After making my Ph.D. thesis proposal in 2011, I went to work full-time in 2012 for Safaba Translation Solutions, a Pittsburgh start-up company offering customized MT systems to corporate clients. From 2015 to 2024, I was an applied scientist at Amazon, working mainly on MT: first on in-house custom systems, then on the general-purpose and publicly available Amazon Translate service, and finally on real-time translation of search queries. I also contributed to a few projects on spelling correction.
See my complete publications list.
If you're curious about what an LTI education looked like between 2005 and 2012 — and my continuing interest in teaching — visit my page of courses taken and taught.
You can connect with me on Github or LinkedIn.
Page created 26 August 2005; last updated 9 September 2026.