>> On Finding Work in Industry

Claudia Brugman, Ph.D.

I went into the industry job market completely unprepared for the differences between academic and non-academic linguistics (and working conditions). Some of the advice below represents things I didn't do right. I now work in the Natural Language Processing group of a company that does information extraction on the biomedical literature.

Besides learning some computational linguistics and programming (as discussed in John Moyne's article), it probably pays to learn a little bit about all the subfields of linguistics, because you never know what kind of work will be asked of you. Learn to work within different theoretical paradigms, but don't worry about being "cutting edge" in all of them. It's unlikely that you're going to have the luxury of using only your favorite theory.

Doing field work in a different language might come in handy, if the work relates to a human user at any point. This is for two reasons: one is that you learn methods for finding out speakers' attitudes about language use, and how to discover preferences in usage, etc. Some sociolinguistics is particularly good background if you get a job that involves an "end-user". When you talk to a prospective employer, you can talk about this experience in terms of providing a more marketable natural-language product to the customer. Secondly, having analytic knowledge of one or more languages that don't resemble English can give you insights into natural-language solutions, even if the language you're working on and in is English. (For instance, I had to construct a syntactic construction in English that has the function of the Japanese "wa"-construction; I might not have thought of it that way if I hadn't known that much about Japanese.) If you're semi-fluent in another language, get more fluent. This could help in two ways: you can do NLP work on that language, and/or you can work on an NLP product (perhaps in and on English) in a country that uses that language.

Maximize what's called "domain knowledge". Much NLP work is in a particular domain of expertise, working with language materials relating to some subject matter (such as molecular genetics or cooking). You can maximize your domain knowledge by becoming slightly more expert at things you already like. If you're interested in cooking, for instance, you could increase your knowledge of cuisines of the world, nutrition, and related areas, in case a job comes up that requires you to do NLP work for an on-line culinary advice column. Better yet, try to find out what domain knowledge is likely to be in demand in NLP fields (peruse the job listings), and do some coursework in that domain while you're in graduate school, or promise to take courses in the domain. (I'm taking a molecular biology course.) Someone with linguistics background and appropriate domain knowledge will be highly sought-after in companies like mine.

Maintained by Sally Morrison.