From Linguistics to Computational Linguistics

Computational LingusticsA couple of undergraduates have written to me recently asking the same, very pertinent question.  Here’s an example we had in recently from Emily:

I am just finishing my undergraduate degree in English Language/Linguistics and I’m considering continuing to postgraduate study in computational linguistics (for example at Sheffield doing MSc in Computer Science with Speech and Language Processing) in a few years. My question was, if I did go on to achieve this masters qualification would this set me up to be able to apply for some of the jobs in this area?

The reason this question is so pertinent is because it touches upon 2 real issues that faces computationally-minded Linguistics graduates in particular: First, the vast majority of work in Computational Linguistics is now happening in software development environments, and software companies tend to value Computer Science degrees highly.  Secondly, the way that the technology world approaches linguistic analysis is predominantly statistical, and linguists tend to be taught rule-based approaches instead.  In fact, some would argue that Computer Scientists and Linguists are a long way apart in how they view language, and that mathematicians are better qualified to work in Computational Linguistics than most linguists.

So with this context in mind, here’s a few pointers I would give Emily, and anyone else in her shoes:

1 Do the Masters or get another career
I’ll go out on a limb here and say it’s not possible for a Linguistics graduate (Bachelors) to be taken seriously for the majority of positions in Computational Linguistics, UNLESS that person already has very well developed programming skills and has completed relevant and interesting internships in Software Development companies.

For the rest of us, if we’re serious about getting into this industry, a Masters in Computational Linguistics or Language Processing or Informatics is an absolute must if you are even remotely serious about moving into this industry.

2 Learn to code, do it now
I have been told that there are plenty of Computational Linguistics PhD graduates who don’t really program.  That’s not a slur, but it’s true that you can get through a postgraduate course of study without really having to code in a way that mirrors what you would experience in an industrial setting.

Whilst most Universities will welcome Linguistics grads with limited or no coding experience, most companies developing language technology won’t.  You don’t need to be world class (I mean, sure, it would help), but you need to know how to build software and the fundamentals of key programming languages.  Python is popular but – in my experience – is rarely enough on its own.  Learning some Java or some C++ and being able to demonstrate that you can use them effectively would give your prospects a serious boost.


3 Get some experience or build something…and then talk about it
Lots of postgraduates come out of academia with a great degree and a bunch of enthusiam.  They then get beaten to the best jobs by people with similar academic profiles and some real-world experience.  In some cases that’s an internship, in other cases people have just built something. 

That sounds scary but actually it’s quite simple – pretty much all of you will build something interesting on your Masters, but most of you won’t tell anyone about it.  If Internships aren’t for you (and I get that), then why not take your Masters project, develop it a bit in your own time, and then get it on-line.  Built a PoS tagger?  Great!  Can I see it?  Developed a prototype ASR? Awesome.  Put a nice UI on it, get it hosted somewhere and tell us about it on your profile.  You won’t believe how powerful this is.

4 Look at the long term
If any of this has made the journey sound difficult, don’t be disheartened.  Natural Language Technology is an emerging industry, where difficult science is applied to very difficult problems with often mixed results.  It’s difficult to make money out of mixed results, and as a result it is still a relatively immature industry in commercial terms. Getting into the industry isn’t straightforward – and in truth it’s expensive (as it’s essentially closed off to non-postgrads) – but the long-term earnings potential here is significant. If you are prepared to be patient and invest in yourself then you’ll have a career at the cutting edge of tech and a very marketable skillset.

I should also say that there are some fantastic research teams at Universities across the EU who are doing some very exciting things, and lots of people mix research work in academia with commercial work in a variety of environments. 

We should stress that the above is merely one view of a broad and diverse industry.  It was written by Nick Gallimore, of the EU’s leading recruiters in the Natural Language Technology space.  You are urged to seek the advice of a range of people when making serious, life-changing decisions.  If you would like to speak to Nick to get his advice on your situation, please drop him an email at


About the author

Nick Gallimore is a classically-trained linguist with a strong interest in Computational Linguistics & Natural Language Processing. He has successfully combined this passion with nearly 10 years’ experience recruiting technology professionals for software companies. The result is an exceptionally well-connected recruiter in the Language Technology space, who comes with the added bonus of knowing what he’s talking about, at least some of the time. 


Leave a Reply