Should I take a break from my job to learn Deep Learning?
Disclaimer - Deep Learning and Machine learning are very hard.
I get it totally, you might be in at a stage where your intellectual needs are not being met by your job and you really want to upskill yourself. This not be only because you need a promotion but also because you really like or are fascinated by the field of artificial intelligence and you want to contribute your share.
You know there is a trend always, always! In 19th century all the intellectuals wanted to be leaders in the field of mathematics. Then as mathematics evolved, the field of theoretical physics thrived! People started moving towards Physics in 20th century and that allowed humanity to make great and rapid progress in the fields of electronics which led us to 21st century. Today most of the great minds are working in the field of Artificial Intelligence.
Mind you, the field of Machine Learning is not new. It has been there since a long period of time. You can read the paper "Neural Networks: What are they and why is everyone so interested in them?" published in 1987. Multi-layer perceptron usage can be dated back to 1960s. There was a stiff competition from support vector machine which was faced by deep neural networks. It was backpropagation which re-ignited the field.
Why am I telling you all this? I am making you aware of the FOMO(fear of missing out) which is around Deep Learning and the various people are cashing it out. It is not just you but even companies are falling in this trap.
The people who succeed in any field spend 10000s of hours honing their skill. So my friend you have to have patience!
1. Start Small
If you are a beginner and looking to start learning deep learning then I'd suggest to try giving small amount of time, say 30 mins in listening to podcasts like TwiML while travelling and start with a basic course about deep learning(from deeplearning.ai) after office, average length of videos in this course is about 15 mins.
2. From noob to beginner
Now you have familiarized yourself with the world of deep learning and you are now sure that you want to continue in this field. It is time for fast.ai mooc now. These lectures are 1.5 hours long and it will require serious efforts from you after office. However, this will make sure that you are now ready to take next step. Read - How not to do fastai course?
3. Start applying for jobs
Now you are comfortable with deep learning but you still are not at a stage where you can leave your job. Now you might be realizing that you still are a lot far from the world where innovators are working. But you know what, it doesn't matter. You are in your own game, it is not a zero sum game either. Start applying for junior data scientist positions from angellist. Remember you are still not a data scientist! So be willing to accept failure if you are not getting selected.
3.5 Start reading research papers.
4. Time to step up the game
a. If you weren't able to get any job then it is time to be serious. It means that you need to put in some serious efforts. You should work hard to demonstrate your skills. You should start putting projects on GitHub and if you can try participating in kaggle. Don't mention your kaggle ranks in Resume if you weren't able to get in top 50. Along with it try to further increase your knowledge in Artificial Intelligence. Start taking up cs229n. It is time to increase your horizon and get complete acquaintance with Machine Learning world as well.
Remember - Keep applying whenever you feel you have increased your skill.
b. If you got a job, great! Start reading some really good books about Artificial Intelligence. "Pattern recognition and machine learning" by Bishop, "The deep learning book" by Ian goodfellow etc.
We never recommend you to leave your job! But we try to aware you that the field of deep learning and machine learning is hard and uncool, you need to have patience here.