Introduction:
I have been working with mentorbruh, a close impact mentorship program for quite sometime now, and had the opportunity to meet a lot and mentor a few data aspirants. Lots of them, come with a prior experiences with internship programs, coaching schools and other type of certification courses and what not. But most of them one thing common.
"Their knowledge from these programs didn't make them employable."
Now when I say that, that sounds like I am exaggerating; and it can't be like that. But in this article, I am going to explain three types of programs, and also will in detail break down that why they don't work fully or partially.
Summary of article:
I am going to take a deep dive into the current famous programs running in india for data science coaching, and will thoroughly explain their flaws and reason of very small to no success rates.
The SMB coaching centers:
The lowest tier in data science mentorship is the small and medium businesses banking on data science coaching. These are 10-500 people sized and most of the times, are run by one main head personality with multiple teachers. Given the structure and low cap of income in such institutes; the quality is the first thing which takes a hit. There are obviously a few 0.1% exceptions; but most of these coaching institutes get 3 things wrong:
(1) the data science knowledge of the professors/mentors are not guaranteed here. Having a 25-30k bound salary in most cases, the professors in these coachings are people with limited exposure to the subject themselves. So the knowledge that goes to students is even less.
(2) There are just too many people in the classes most of the time. Generally coachings run a minimum of 20-30 people batches; with 1-1.5 hr long classes; where there is already a lot of "material" to teach on. This classroom structure is a complete no-go for subjects like data science; and given that there is almost no one to one timing in such system, there is absolutely no guarantee that the student understands the stuff.
(3) Lack of quality checks is the last nail in the coffin of these systems. The system in such a organization mostly involves a fixed number of courses, few assignments and then ends up with a certificate. The problem with this is that while assignments are mostly subject to copy paste and other obvious problems; there is no actual check that whether the students learn or not.
Most of the students whom I have interviewed, lacked basic theoretical understanding, as well as didn't know what they did and why they did. This proves the problem with these category of institutes.
Online project internships:
This is the second type of programs a lot of data aspirants go through. Online project internships, with the likes of the sparks foundation program, has more success rates than your typical coaching institutes. These programs, come with a fixed set of projects and a fees, and gathers a large number of participants in online communities and make them do the projects within a stipulated time. The good thing with this is that the participants are forced to do hands-on; and are often required to share the results in social media platform like linkedin; making their work a bit more valuable than totally unsupervised projects.
Then why am I criticizing this format? The criticism of this format comes again from the above para about coaching.
(1) In a online project internship format, once again, there is little to no contact with actual mentor. The purpose of doing a good project for data aspirants is not writing some fancy codes and showing results, but getting a solid background for the projects so that next time they can do it on their own. The major part of this comes from the mentor for the project. And this is the exact part which is totally missed from the online project internship.
(2) There are, again, too many students doing the same thing. In data science industry, people are given value on two things: (1) in-depth skills and (2) creativity. The problem with this method is that projects suggested in the online internships are often very shallow, and gives little exposure to the subject in hand. And the fact that so many students do the same thing, guarantees the lack of individuality and uniqueness of the program.
(3) Finally there is a lack of checking the quality. Without effective exit exams, there is no guarantee on how much the student actually learns from the program.
Most of the students coming out of these projects, are however a bit better than coaching center students, and people with high self-driving factors, even learn something from this. But as I analyzed, there is a lack of proper structure which doesn't ensure the quality of education in this process.
The costly online certification programs
Then there is the last category. Online certifications are not too shabby as the last two ones, but there are still a lot of holes in this ground as well which slips a lot of aspirants to lose both money and time.The main issues are:
(1) Many certificates are, in plain terms, worthless. If you are in-experienced and as beginners mostly are, they fall prey to these. After 6-8 months, and a few lakh rupees out of their pockets, they end up with a degree/certificate, which means nothing in the real interviews.
(2) Among the good certificates, a few are really really costly. What I mean by that is, there is already a lot of quality contents out there for free; which is enough to train a data aspirant upto medium level data scientists. The part which is not available are business acumen and industrial experience, which is anyway impossible to teach through a course. Hence if you are paying something around 3-6 lakh for a degree; well you are paying upto 10x more price for what you are getting.
(3) Other than a few good certificate and degree courses, like the ones from upgrad; most don't offer any credible placement offers as well. While placement offering is not a main thing in most coachings, it is often added as a part of the packages which the students buy when they sign up for the courses. But seldom these turn out to be that good.
Conclusion:
In conclusion, most of the current famous structured programs, are the ones to be taken with more than a pinch of salt. If you are in one of the above programs, chances are, unfortunately, you are not going to get much better at the end than the beginning if you don't try to learn on your own. There are a few good programs like Alpha ai or my own mentorbruh; where the hands-on is being focused, as well as students are subjected to present their work to a number of people to ensure the growth. The future of data science training is with p2p, closely mentored programs with full hands on projects. If you want to join such a program, Read about mentorbruh here; and you can contact me on shyambhu20@gmail.com for further discussion.
Thanks for reading! stay tuned and subscribe to my blog for getting awesome data science posts like this to your gmail feed every week.
Comments
Post a Comment