I want to become a language engineer. But how?
You're a linguist and want to specialise in Natural Language Processing. Good news: we're in the same boat. Not-so-good news: it's going to take some time and you'll have to put a good effort into it.
So the big question is: How do I train to become a language engineer?
In the last few months, I've talked to quite a few people who work in tech, and done extensive research on the Internet. What I have learnt is that the route to NLP comprises three major steps:
1 - Learn to code in a language of your choice (possibly Python, given that it is widely used in AI and especially NLP);
2- Learn Machine and Deep Learning;
3- Learn Natural Language Processing.
I have tried to learn all these by enrolling on a MSc in Computer Science with Artificial Intelligence from the University of York. However, I found the course too theoretical and not quite as practical as I was hoping it to be. It appears that recruiters in industry are not really after degrees: they look for experience and practical skills, which is why I applied for a leave of absence from my MSc and decided to take hands-on courses instead.
How do I learn to code?
There are many resources out there, and some are of course less expensive than others. The final choice will depend on your personal finances and also, on your learning preferences, and on the amount of time that you can spend on your learning. For me, it's going to be part-time evening learning, as I'm currently working full-time on a research project at the University of Cambridge, and I have two young children at home.
My suggestion is to do your research and see what's best for you. Free or rather cheap options are Coursera, the Python Institute, Codecademy, HackerRank, Code Gym, Udemy, single courses at FutureLearn, single courses at SimpliLearn. On the most expensive side, you have the choice between Udacity, EdX, professional courses at FutureLearn, degrees at SimpliLearn. And much more...
Take into consideration that you can learn a programming language rather quickly, but then you'll need to learn how to use the libraries related to your field of choice and also, quite a lot of practice. My suggestion is to forecast a 2-3 month learning process for this stage.
In the next post, I'll share my final choices as to how to learn to code. I started last week and I must say I'm very pleased!
How do I learn Machine Learning and Deep Learning?
I've done quite a lot of research on the matter, and my favourite resources are the following. I'll list them from least to most expensive:
- The IBM Machine Learning professional certificate on Coursera;
- The IBM Deep Learning professional certificate from EdX;
- The Deep Learning nanodegree from Udacity;
- The Advanced Certificate Program in Machine Learning and Natural Language Processing from UpGrad (bonus point: it's already got a module on NLP, which can save you some time later);
- The Ai.Core Machine Learning engineering bootcamp (it comes with a price tag of £6500 but provides live, interactive learning and career support).
These of course are nothing but my favourite courses, and I'm only mentioning them to help you out in your research. Take these as a starting point and then try to figure out what's best for you. You'll see there aro so many resources out there you'll often feel overwhelmed.
All of these courses are quite advanced and provide hands-on experience and the chance to work on real-world projects. Prepare to spend 4-5 months on them.
How do I learn Natural Language Processing?
There are quite a few free resources on the Internet, but my heart is set on the following two courses:
- The Natural Language Processing nanodegree from Udacity.
In a later post, I'll go over these thoroughly, and discuss the pros and cons of both. For now, take into consideration that the DeepLearning.ai course is way cheaper than the one from Udacity, although Udacity can offer attractive discounts (of up to 75% of the original price tag) if you're a student. Both last about 3 months.
To sum up: How long will it take to up skill and be industry-ready and how much is this going to cost me?
You'll need about 9 to 12 months to up skill properly if you're a full-time professional. Cost wise, the price will range from a couple hundred pounds to up to £8000+ if you decide to take the Ai.Core bootcamp and two Udacity nanodegrees. Eventually, the choice is yours.
Switching to a new career is a life-changing choice, so my best piece of advice would be to think about it thoroughly and don't rush anything. A good night sleep is often needed before making important decisions. Give yourself two or more nights to think about this, no one is running after you.
If you follow me in this journey, you'll see a real example of career switch from post-doctoral researcher in linguistics to language engineer. I'll share everything: from insights on the courses that I will be taking, to pieces of advice, to my struggles and victories. And then I'll document my job search too, of course.
So what do you think? Have you done any of the above courses and are ready to provide some feedback? Or maybe you know other courses that seem more relevant to you? Do you have the same career ambitions as myself and want to join me in this learning adventure?