recategorized by
3,499 views
6 votes
6 votes
Is getting into iit's necessary to become good at machine learning and data science or there are other ways to make a career in it ? If iit is necessary then what is cutoff for selecting this stream. if someone could write a post on this topic or answer in comments then it will be very helpful for me in planning accordingly.
recategorized by

2 Answers

Best answer
7 votes
7 votes

Great question. Even I was faced with this dilemma. After numerous emails to professors and advices from them this is what I got.

Machine learning is really a nice field to go forth. You could definitely learn much of it yourself and probably get a job in it but it isn't fun doing everything yourself alone.

Sitting inside a lab with resources, amazing talented individuals around you and most importantly the teachers who will guide you to excel and climb, won't be with you if you sit and start yourself. It's not a necessity that you join an IIT or perhaps IISc ( has 4 ML Labs with a supercomputer with 126720 Cuda cores with 2.75TB of Graphics RAM, just saying. ) to actually do some good machine learning and AI work. Trust me you'll miss something and being in academia is swag. :) ( publishing a paper is so frustratingly rewarding )

You could practice at Kaggle, do some projects and slowly improve by own if you got the motivation and then get a nice job as a data scientist but most good companies look for qualifications.

Personally, I advise you to prepare for GATE and get into IISc or IIT's because after you get in, you get the exposure, the atmosphere and most importantly you'll fail for people are always better than you. You'll fail to stand up. When you're at your own you won't fall to stand. 

I'm probably clear for the last line speaks your answer. 

All the best. I seriously do believe you'll do good. 
Ankit.

P.S:  Fun video about a neural network learning to generate voice here. 

edited by
0 votes
0 votes
You Don't need to go for GATE or IITies. To become an awesome data scientist or a ML expert, you might need in-depth knowledge of machine learning, artificial intelligence, parallel computing in clusters, and a good deal of core statistics.

These are very specific areas and its a fraction of the syllabi covered in ME / MTech. Also, the depth that is required to work at the cutting edge of these domains are not provided in the masters level. You might have to pursue a Ph.D to gain true expertise in these areas.

One more thing, data scientists need to ask the right questions and formulate the right hypothesis to prove or disprove, apart from making it easily digestible for other departments of a company. The business sense and the communication skills are extremely important - you will not get that in academics.

So my suggestion is take some focussed MOOCs like Andrew NG's Machine Learning course on Coursera to begin with, do small projects by making use of Python's data science libraries, participate in Kaggle competitions, learn R, learn some core statistics.

If you are interested in data engineering, learn the big data tools like Hadoop, Mapreduce, Pig, Hive, etc.

No MTech course in India will make you industry ready. So invest your own time in building right skillset besides doing your day job, and then look for the right opportunities in data science.

Good Luck!!!

Related questions

1 votes
1 votes
0 answers
4
Aegon asked Apr 9, 2018
826 views
Recently, IITH has introduced machine learning as a choice beside general CS. As its newly introduced is it better to give it a first priority than General CS if someone...