The Gateway to Computer Science Excellence
+39 votes
Which IIT's take admission through written Test and Interviews,Upto what Rank They call and  
Eg Major Subjects To Focus and Type of Questions asked.
in IISc/IITs by Loyal (8.2k points) | 8k views

2 Answers

+35 votes
Best answer
For " Mtech CSE (TA) " Admissions in IIT , All IIT's except IITD and IITK give direct admission based on Gate Score.

IITD Admission Procedure :

IITD gives direct admission based on gate score and btech percentage .Direct admission is  myth here (rank 66 didn't get direct admission last year and those who got ,left the seat and went to IITB/IISC). IITD calls upto 250 rank(approx) for written/interview. Here also they had percentage criteria of 67.5(or may be 70) last year.First there was one round of written(3 questions were asked related to algorithms and mathematics).Time duration was 1 hour approx and it was then followed by one interview based on written exam questions.After then they shortlisted some students and then there was another round of interview.Finally some candidates were offered admission there and some were kept in waiting list (upto 20 ,which was completely cleared later) .Number of seats are 35 there.

IITK Admission Procedure :

IITK gave 33 direct admissions based on gate score + btech percentage.Last rank who got IITK direct was 103. They called upto 550 rank(approx) students for written/interview.In written there were 30 questions (25 multiple choice questions and 5 without options).All questions were of 3 marks.Multiple choice questions had negative marking also(-1 mark).Time duration was 2.5 hours and the total weightage was 90 marks. After written they offered admission to 19 students(who secured >51 marks) and shortlisted 47 students (who secured >42 marks).After taking Interviews , finally they offered admissions to 19 more students.So in total the intake was 33 + 19 +19 .They didnt put any waiting list.Out of first 33 students not more than 8 students are here in IITK with me . And the total current strength here is 41.


Prepare maths and algorithms for sure. In every interview u will face questions from these subjects.Apart from that also , don't leave any subject. Anything can be asked.Nothing is predictable.Good Luck to all :)
by (325 points)
selected by
Thanks Gaurav, Regarding maths and algorithms same approach as for gate or more depth required ?
IITs which does not consider percentage for admission

@Anurag_s ... Same approach as for gate is sufficent , but It would be better if u study " theorem with proofs" of discrete maths also because a maths professor would love to ask proofs.Also do solve some puzzles(both mathematical and technical) .You can find interview puzzles here also: .

Good luck .

Sorry @khamer . All IIT's have some minimum percentage criteria.You need to go through their website individually to get to know minimum percentage criteria.
Hey..About the test they conduct, is it subjective or objective like GATE??
In my time ,

IITD - all 3 subjective

IITK - 25 objective and 5 fill in the blank (can be considered subjective) :)
thanks :)
Is anything have done to save those seats which are left by direct admit students?
Last year COAP was used for most admissions. So, vacant seats were quite negligible.
+43 votes

Earlier at least half of the IITs had interviews for M.Tech. Now most are having direct admits. 

For actual cutoff range please see the respective tab at bottom:

IISc. Offers direct admit for ME (same as M.Tech. elsewhere) in CSA, SSA, and DESE departments. CSA is Computer Science department, SSA is a joint one of CSA and EE depts and DESE is more on electronic side (embedded). Then there is M.Tech. in SERC which is the super computing department and works on many system side subjects. This has interview. Also, all these departments take interviews for MSc. (same as MS elsewhere).

IITB: M.Tech TA is the normal 2 year M.Tech. and is direct admit. There is also M.Tech. 3 years which has interview. This can be considered like MS elsewhere. 

IITD, IITK: Offers direct admit for select few students (who have high CGPA and GATE score) and for others interview is there. 

IITKgp, IITG, IITM, IITH: Direct admits for M.Tech. interviews for MS.

More information including interview experience you can get from

For interviews there might be a screening round where GATE level questions are asked. Sometimes it is more on aptitude side and last year some had plenty of geometry questions also. But areas deep into CN, CO etc. are usually not asked and maths, algorithm, operating system etc. are more important. 

Now, for interviews your profile has a huge impact. I say about research interviews only as for M.Tech, it is usually for namesake.

  • Say if you are a university topper (a university topper in PG has research admit granted due to INSPIRE scholarship) or with high CGPA, NTSE or similar scholarship winner, some other achievements etc. you have a clear advantage.
  • Another criteria is research publication- if you have published a work somewhere, it shows that you have research interest- even if work is not great, and professors value it. So, if you are a final year student get your project published- it is not very difficult. 
  • Be smart in some areas. For GATE one need not be smart in any areas and just basic level knowledge is enough. But for demonstrating research interest one need to show real knowledge - mostly outside text book. This is easy for someone who work in a related area on company- say like those working in Intel and CO area, Cisco and Network area etc. Otherwise one can read the recent publications of the professors of the IITs you are going and get an idea of their work. Reading papers might not be easy but it would help. 
  • Selecting area of interest: This is quite tricky because unless experienced you can't really know your strength as GATE is a very basic level exam. In GATE you are studying like 15 subjects say at 5% depth and now you are going to take 1-2 among them and going to study at 50-80% depth. So, you shouldn't decide this based on your scoring in GATE or B.Tech. For example, in research in DBMS one mostly works on query optimization and not SQL queries as in B.Tech. In Compilers also one works on code optimizations and mostly not on parsing. Most people select algorithms and data structure- but these are mostly theoretical area and recommended only for those who want to be in academia. System side like OS, Compilers, CO etc can get you job in Intel, MS, Google, Nvidia etc. and Intelligence side like Machine Learning can get you jobs in ecommerce giants like Amazon as well as Google, MS, FB etc. Those who are good at GATE questions in CO, OS can safely take system side. Those who are good in probability can take Intelligence area. 

If your profile is normal, I'm no way saying that you cannot be selected- many people do. My only point is that if you add something extra you do have a better chance. A good performance in interview gives selection most of the time.

by Veteran (430k points)
edited by
Hi Arjun Sir,

Taking systems side and gaining hands on experience in systems programming can be a plus point for interviews, but as the syllabus is vast, one needs time for it.
What do you advise, should I dedicate my time to it, or instead focus more on algorithms and coding.

Both are not easy. So, it depends on what you find easy. Algorithm/small coding is unavoidable as it might be needed in written test. But as a final year student (or even passed out) I would advise to get your project published as it has a LOT of weightage for research admission. Most probably once you clear written you will be selected based on your profile and your research interest and not on the basis of the number of questions you answer. 

IISc calls ME as MTech & MSc(Engg) as MTech(Research) According to
Good change of names :) But they don't restrict M.Tech. Research to 3 years as IITB rt?
@Arjun, what are the main subjects that we studied for gate will be useful in intelligent systems while pursuing mtech Reasearch. Can u tell how to decide  area bw systems and intelligent systems with some more details if possible .
@Anurag_s for Intelligent system you must prepare one of the Linear Algebra or Probability (Complete, both discrete as well as continious).

Check this one ->
yes akash i have read it on your blog, also read on iisc website.

Actually i want to know more about this area that it suits me or not because if i clear interview i have to remain in same area and cannot change.I like probability but i also like subjects such as databases .
@Anurag Yes, it is very difficult to decide a research area based on B.Tech. knowledge.

1. Compilers- in B.Tech. it is mostly about parsing while in Masters it will be mostly about optimization.

2. Databases- in B.Tech. it is mostly theory and SQL while in Masters it will be mostly Query optimization.

3. Algorithms, TOC, Graph theory etc. are more closely related in Masters.

4. Most subject will require probability or linear algebra except system ones. But even system subjects might require these as there are works on machine learning even in architectures. In any good place basics of set theory can not be skipped.

5. AFAIK intelligence area in IISc. requires linear algebra and probability and if you are good enough with GATE portions it should be fine. No need of advance stuffs as for research a good and clear basic knowledge is more important. And it is the most in demand area in IISc. - so if you like a bit of maths you will never regret taking..
Hi Arjun Sir,

What type of questions or level of questions might be asked if research area is Machine Learning, as I followed Coursera course on ML by Prof. Andrew Ng.
I expect questions on basic probability but don't know more. I advise you to just read the research paper on the IIT/IISc. faculty where you are going for interview.
Thank you Arjun Sir
Is CGPA a very big factor?

Is CGPA>8 good enough?
@Arjun Sir,

For ML what extent having avg background in competitive programming and some good points at codechef will help ?

About Linear Algebra...i honestly don't have any visualization like what is eigen value in n dimensional matrix (n >3) is even possible in real does it look like..n such things

So how we r supposed to get visualizations n a clear understanding of these things without the help of Profs who ask these things in interview itself ? Aren't we supposed to learn such things after admission in IITs ? or such mathematical  understandings r pre-requisite for iits ?

pls share any source for understanding mathematics in real..
@Kaushik not a big factor, though being a college topper might help.

@Nitesh To the extend that you can stand out from others appearing in interviews. For understanding mathematics well. I also do not know good sources. Both linear algebra and machine learning are not my areas so I cannot help here. @Suraj might help.
@Nitesh ML has nothing to do with competitive programming.

One should know basics of Matrices and probability.

Best source for Linear Algebra is videos of Gilbert Strang.

At max, You just need to check out videos and how different algorithms are applicable for some of standard problems in ML.

@Nitesh , I would say to watch Gilbert strang videos for linear algebra.

and focus on very basic things (rank of a matrix,  maximum number of non-zero eigen values, etc.) . Most asked ML questions (even in job interviews) are from linear algebra.

Gilbert Strang video lecture series on linear algebra is best to brush up your LA concepts.  You will also solve your visualization problem to some extent. For ML side, you should have a strong hold on Linear Algebra concepts, you might clear the interviews with basics but for taking courses here in IISC, you need a strong hold on your LA concepts. And Strang lecture series is an awesome tool to start with.
Also, In IISc interviews they don't expect you to know ML. But they definitely expect you  to know your LA and Probability well.
Quick search syntax
tags tag:apple
author user:martin
title title:apple
content content:apple
exclude -tag:apple
force match +apple
views views:100
score score:10
answers answers:2
is accepted isaccepted:true
is closed isclosed:true
50,741 questions
57,242 answers
104,604 users