I'm applying to grad school this year, but there is a side of me that wonders if doing udacity courses and side projects would be enough to get a job doing challenging work. What is it like being the machine learning engineer? And the scientist?
Tech Industry
Yesterday
2487
Quitting this Slave life
Working Parents
Yesterday
770
What do you think is wrong with a kid who got rejected by 9 colleges?
Tech Industry
Yesterday
1984
The end of Backdoor Roth?!
Software Engineering Career
7h
292
Salesforce SMTS vs Google L4
India
Yesterday
838
Modi is a legend, will be remembered for centuries to come
Getting a break in ML without a degree is fairly hard but still quite doable as long as you want to be in applied side of things (as opposed to research side). First, you will need LOT of self study. Not just ML algos but also calculus, advanced linear algebra, probability etc. This is very hard to do unless you are really enrolled in courses and obligated to do homework. I would estimate 90% of people don't have self-discipline to do this without getting enrolled in classes. Second, you will need to do more than one big side projects and participate in competitions like at kaggle at least half dozen times a year. In my estimate, it roughly takes 2-3 year starting from BS degree to get to a point where you can do as good a job in developing models as pros can. Then there is whole bunch of areas such as distributed learning, deep learning and so on - each for which need another 1-2 years of effort to specialize. One alternative of doing side projects is to find a team that allows for learning skills on the job. There are not too many teams like that but the good news is that demand currently far outstrips the supply. There there are often cases where team already has bunch of experienced ML PhDs and they want someone with great track record and who is passionate about learning.