From web programmer to data scientist. Optimal vector?

Question

Programming

Good afternoon, dear colleagues. At the moment I am a web developer in the java stack, but I have always leaned towards the field of data analysis, machine learning, math. disciplines, etc. I got tired of web development after 7 years. I decided to retrain as a data scientist. My current position:
Mathematics. He graduated from graduate school, but did not defend in the direction of 13/05/18. There are gaps in knowledge, but mat. analysis, lin. algebra, numerical methods, discrete mathematics, etc. I can pull myself up.
Programming. There are no problems with python, rock, or R.
Machine learning. I can study Vorontsov's lectures myself.
The following questions arise:
1) Maybe not be so presumptuous and enter, say, a fur-mat. NSU? (second higher education) Or finish only a master's degree?
2) Leave NSU alone and just enter the ShAD, take courses on Kurser, participate in competitions on Kaggle?
What would you advise to save time and get a job as a data scientist?

Answer 1

This would be the best option.
You already have a mathematical education, so there is already such a checkmark in the resume. You have programming. What does it take to become a data engineer? - Get a job. So send your resume, go to interviews and see how things are going.
If you are not hired, then look at why and already decide how to fill in the gaps. If you were hired, then you became an engineer, and then it’s a matter of practice.

Answer 2

0. It’s worth asking those who have gone through the path
1. Nobody bothers (like) going to lectures in universities without admission
2. Just relearning mathematics that IMO forgot is not worth it
3. Practice from the very beginning
The tags to the question have answers which books you will need to learn when you really need to pull up mathematics

Answer 3

ShAD will definitely not be superfluous :) There you will gain both knowledge, and acquaintances, and a reputation ....
Another simple observation from life: all my acquaintances from ShAD are hunted incredibly actively, and they hunt for such positions as Head Of Analysis, etc. .

Answer 4

Data Scientist is a vague wording for me.
Data analysis !== Machine learning. Here you need to decide what you want.
Data analysis - work in excel, database, python.
Machine learning - development of models and algorithms.
The following works in both cases.
1) Recruit pieces of 10 vacancies for the desired position
2) Make a list of skills. If the requirement is in many vacancies, then it is important, if in one, we omit it
3) From the list of the most important skills, we pump over the first half. It doesn't matter where, as mentioned above, the main thing is practice
4) We pass an interview for a real assessment of knowledge. Do you understand that they can take on a small position and with a lower salary than you have now?
Pass the ShAD - super.
Read the news and trends in the world of data analysis: datatau.com
Watch YouTube talks with PyData: https://www.youtube.com/user/PyDataTV