IIIT-H Lateral Entry CSD FAQ

Note: Most of the questions answered here are CSD specific.

Questions about the Admission Process

Q: What is Lateral Entry? How can I apply?

A: Lateral entry is a program offered by IIIT-H, which admits students into their dual degree
programs in Computer Science and Engineering (CSD) and Electronics and Communication
Engineering (ECD). All details regarding the application process can be found on the
admissions website.

Q: How many people are selected through the LE admission?

A: The upper bound on the number of students, at least till 2018, was 15 for CSD and 15 for
ECD. However, keep in mind that the no. of students that are actually selected can vary.

Q: Will I join in the second year or the third year?

A: If you are offered an admission to IIIT through the lateral entry scheme you will be joining in
second year i.e. you will sit along with 2nd year B.Tech students of your respective branches.
However, some of your courses could be different from the regular second years to account for
what you studied in your previous institution. CSD students are sometimes allowed to take
electives in their 3rd and 4th semesters if they have already covered a course in the previous
institute.

Q: How many years is the program?

A: The residency requirement prescribes a minimum of 4 years. However, the actual time of
graduation varies depending on how long it takes to complete your master’s thesis (and publish
your research).

Q: Are LEs treated differently during placements/higher studies applications?

A: In all matters, where the college is involved (such as placements), LEs are treated equally to
any other dual-degree student.

Questions about change in dates

Q: The exam/interview dates are clashing with my semester exams? What should I do?

A: It is advised that you send an email to [email protected] about the same, however, in
the past, no changes have been made in the dates.

Questions on mark sheets/GPA

Q: What if I don’t have the mark-sheets up to semester 3 while filling the form?

A: There have been students in the past who have had this problem. Simply upload the
mark-sheets that you currently have.

Q: What is the GPA requirement?

A: For 2020 Lateral Entry Admissions, there was NO restriction on the CGPA, implying that a
candidate with any CGPA can apply for the entrance exam. However the candidate should
not have any backlog.
Until 2019, the institute required that the candidate had a minimum of 8 CGPA . Restrictions on
CGPA for LEEE 2021 are not yet clear as of now (January 2021).

Q: How is the Cumulative GPA calculated?

Σ(𝑔𝑟𝑎𝑑𝑒 𝑖𝑛 𝑐𝑜𝑢𝑟𝑠𝑒 * 𝑐𝑟𝑒𝑑𝑖𝑡 𝑜𝑓 𝑐𝑜𝑢𝑟𝑠𝑒)

A:
Σ(𝑐𝑟𝑒𝑑𝑖𝑡𝑠 𝑖𝑛 𝑠𝑒𝑚𝑒𝑠𝑡𝑒𝑟𝑠)

Note: It is not the average of GPA of each semester.

Questions on eligibility

Q: I am from EEE/Electrical and not ECE, can I apply for ECD/CSD?

A: Yes, you can apply to both ECD and CSD as per the latest rules. But keep in mind you need
to clear the exam and interview of the program you apply for.

Q: I am from IT/ICT/SE and not CSE, can I apply for ECD/CSD?

A: Yes, you can apply to both ECD and CSD as per the latest rules. But keep in mind you need
to clear the exam and interview of the program you apply for.

Q: I am from an integrated BTech/BE + Masters program, can I apply for ECD/CSD?

A: If you are from an integrated BTech/BE + Masters program, only your BTech program would
be sufficient to decide if you can apply for ECD/CSD. Refer to the above two questions.

Q: I am finishing my 3rd year, am I eligible to apply?

A: No. As mentioned on the website, only the students who will be completing their 2nd year by
July 2021 will be eligible to apply.

Questions on online exam/reference material

Q: What is the format of the online exam?

A: The format was changed from subjective to multiple choice in 2018. There will be MCQ,
single correct questions from different aspects of the syllabus.

Q: Which books/reference materials are suggested for the online exam?

A: There are as such no specific books to crack the online exam. Generally, just following the
prescribed books of your college course should be sufficient.
But here are some standard books for each section:
C Programming:
Recommended reading - The C Programming Language by Brian Kernighan and Dennis
Ritchie (for pointers), Let Us C
There are several online resources to learn, practice questions to test your
understanding.
Data Structures:
Recommended reading - Introduction to Algorithms by Charles E. Leiserson, Clifford
Stein, Ronald Rivest, and Thomas H. Cormen (CLRS)
Discrete Mathematics:
Recommended reading - Discrete mathematics and its applications by Kenneth H.
Rosen.
Digital Logic + Computer Organization:
 Recommended reading - Digital Design by Morris Mano, Computer Organization and
Design by David A Patterson John L Hennessy and John L. Hennessy
General Aptitude: No books required.

Q: How is the difficulty level of the exam OR What kind of questions are asked in the
exam?

A: The level of questions was quite different for each section.

C Programming: Questions were mostly related to either debugging or finding the
expected output. Focus on pointers. The questions asked on pointers were of a good
level and required a good knowledge of pointers to solve.
Data Structures: Some of the questions were straightforward. Some questions required
that you find the state after several operations.
You should be familiar with linked lists, arrays, queues, stacks, heaps, etc. There were
1-2 questions based on advanced data structures like AVL/Red-Black Trees.
Implementation is not required, although it is beneficial to know.
Discrete Mathematics: The questions from this section were time-consuming. The
questions were related to combinatorics (P&C), graphs, trees, the colouring of graphs,
minimum spanning trees, etc. Some of the questions were non-intuitive and required
actually proving some stuff to get the answer. You need to be familiar with at least the
terms used in Discrete Math.
Digital Logic + Computer Organization: There were very few questions about pure
digital logic (like flip flops or latches). Most of the questions were CO-related such as
adder circuits and instruction fetch architecture.
General Aptitude: Questions can be based on Logical Reasoning, Pattern finding, etc.
There were very few questions from this section and they were quite easy.

Note: This answer was specific to last year’s (2018) exam. The pattern/difficulty of the exam
may vary from other past exams.

Q: What are the cutoff marks to clear the online exam?

A: In 2018 it was around 32/75 for ECD and 37/75 for CSD. In the years 2019,2020 the students
were only told if they were invited to the interview. So it's hard to tell what's the cutoff since the
candidates weren’t even provided with answer keys.

Questions on “Project/Internship/Research Experience”

Q: Do I have to have a project/internship/research to apply for LE?

A: No, it is not necessary to have a project/internship/research. Many of the students who got
selected did not have any of these. However, having these do show that you’re genuinely
interested in research, which may give you an edge.
Q: Should I do a project/internship to improve my chances of getting in?

A: If you’re doing these solely for the purpose of getting in, then no. Keep in mind that any
project you mention on your application might be a talking point in your interview. Unless you’re
confident enough about your work to answer questions during your interview, do not mention it.
There were students who had a project/internship in their resume but did not get selected.

Questions on the interview

Q: What kind of questions are asked in the interview?

A: The questions asked in your interview, as far as what we have noticed, tend to be related to
what you studied in your 4th semester. Of course, this is not always the case. You can always
guide the panel towards what you want to talk about.
For example - several candidates were asked what their favorite subject was and were
questioned on that.
Anything mentioned in your application can be a talking point. For example - an online course
you did, a project/internship you did, the research interest that you mentioned etc.
Apart from this, there are general questions like “Why IIITH?”, “Why research?”, “What is your
research interest?” etc. Make sure you do your research about the institute to be able to answer
these.

Q: What are some general interview tips?

A: The profs on the panel are your friends, they will try hard to guide you to come up with your
own answers. But do not expect them to solve the questions for you. Ask as many doubts as
you can.
They want to see how you think. Feel free to use the whiteboard to write out your thought
process and explain what you’re doing at every step. Keeping your thoughts to yourself doesn’t
help anyone.
Take any opportunity you can to show them that you can write code. The curriculum here
focuses heavily on programming. If they see you can write code quickly and under pressure, it’s
a point in your favor.
Bring a CV/Resume along with your certificates. They may or may not have time to go
through all your certificates. In some cases, they only glanced at the certificates but spent time
reading the Resume.
Interview experiences:

From Students Selected in 2018 (CSD)

Interview Experience 1:

I was the second person to be called in. There were 4 professors. I somehow managed to drop
the folder containing my certificates on the table, so I guess you could say my interview started
off on an amazing note /s.
I1: *General questions like name, origin, college, etc.*
Me: *Answered the questions*
I1: Have you had a course on Discrete Math?
Me: Unfortunately, no. Our college offers it in the next semester.
I1: That’s surprising, what courses did you have in your previous semester then?
Me: Computer Networks, Operating Systems, Microprocessors... (I conveniently skipped
Probability because I was not confident about that subject)
I1: Ok forget that what is your favorite course?
Me: I would have to say Algorithms.
I1: Algorithms? Ok, so how can I know if an algorithm is running properly?
Me: You would have provided a proof of correctness.
I1: Yes, how would you prove an algorithm is correct?
Me: Generally, if there is an inductive structure we can do a proof by induction. Otherwise, there
are certain properties that we can prove which can, in turn, prove the correctness of our overall
algorithm. For example, Kruskal’s algorithm can be proved by the Cut property. (I1 cut me off
before I started explaining).
I1: Let's not go into graphs, let's keep it simple. Say I need to find the factorial of a number, how
would I prove the algorithm is correct.
Me: (This problem pretty much stumped me because it was so abstract). *I explained the
inductive structure with a base case being fact(0) = 1 and verbally proved it by induction.*
I1: Fantastic, but what if I write it iteratively instead of recursively?
Me: Sir, I don’t think there will any difference because the inductive structure still holds.
I1: Are you sure? I don’t think its the same. You can use the board if you wish. (If they suggest
using the board, they want to see your thought process. Grab the opportunity to write
code/proofs).
Me: *I wrote the iterative code to compute factorial.* (I had no idea how to proceed further).
I1: What is the loop invariant in this code? (I1 was trying to guide me through the proof, but I
had never heard of the term loop invariant before).
Me: Sir, I am not sure what you mean by loop invariant.
I1: You don’t know what a loop invariant is!? You have done model checking right? What is an
invariant? (I am guessing he saw the model checking course I did on my Resume).
Me: It is basically a property that holds throughout the execution of the program.
I1: Good, now what is that property here?
Me: (Totally blank) The variable i is always smaller than equal to n?
I1: That is obvious! That's like saying the sun rises in the east! Tell me something more
interesting. (Those are the exact words he said. I pretty much lost hope at this point. Fortunately
for me, I1 got a phone call which he walked out of the room to take.)
I2: Ok, forget that. You’ve done Operating Systems, right? Say I have a linked list of elements.
Now I have two processes that are both running insertion sort on this linked list. What will
happen?
Me: Sir, in case there is no mutual exclusion, there is a chance for a race condition to occur. (In
retrospect I feel pretty stupid for saying this. It's like saying the sun rises in the east. Again).
I3: Take a random linked list. (I wrote some random numbers). Now tell me what will happen if I
insert 5 into this.
Me: *I had started to work out the insertion when I2 interrupted*
I2: How would you swap two elements?
Me: *I wrote the code for swapping the contents of the nodes*
I2: Now do you see what will happen?
Me: (Realization about why they were making me do this finally dawned on me) *I answered the
question confidently*
I2: (Seemed to be satisfied) You can leave. *Hands me back my folder*

Interview Experience 2:

How did I prepare for the exam?

To be honest, I really did not. I did not take this exam seriously and was preparing for my
end-sem exams which were just two days after the online exam. I had pretty good programming
and mathematics background, so I was confident for sections like C Programming, Discrete
Mathematics, Data Structures, and Algorithms. I remembered some stuff from my course Digital
Logic. For Computer Organization, I didn’t bother studying as this section contains a lot of topics
and it would be time-consuming, taking my end-sems into consideration. On the day of the
exam, I went briefly through pointers from the book “The C Programming Language by Brian
Kernighan and Dennis Ritchie”, as I was weak in pointers. I consider this decision as a wise one
as I was able to solve all the questions on pointers. I didn't study anything other than pointers.

How did I prepare for the interview?

Again, as I said, I was confident in sections like C Programming, Discrete Mathematics, Data
Structures, and Algorithms, I skipped these sections completely. Although, before my interview, I
went through all the sorting algorithms. I focused on preparing for the subjects that I had
recently studied in my 4th sem. There were two relevant courses namely probability and
systems software. Once again a wise decision, I was asked a lot of questions on probability and
preparing for it helped a lot. I focused more on systems software course as it had a lot of
concepts. As this course is somewhat similar to the course operating system, I referred to the
book Operating System Concepts and went through stuff like processes, threading, process
synchronization, deadlocks, etc. I was not asked any question relating to this course. I also
prepared for mathematics courses like group theory, linear algebra, etc as they were a part of
my curriculum.
I briefly went through some machine learning concepts which turned out to be unnecessary.
Note: You are not asked about anything that was not in your curriculum, in the interview.

There were 4 panelists, I1, I2, I3, and I4. I didn’t really expect this.
As soon as I went in, I was asked for my mark sheets/certificates and whichever documents I
had brought, by I1.
I3: So you are from Dhirubhai Ambani Institute…?
Me: Yes.
I2: What are your research interests?
Me: I told them that I am really interested in working on something that includes mathematics
heavily. Also, I said I find projects related to computer vision interesting. Moreover, I had gone
through the labs of IIIT earlier, so I mentioned two of them which were CSTAR and CVIT. They
were happy to know that I know about these labs.
I3: Okay fine. Take this marker and go near the board. You are given a tree with n nodes. How
do you find the distance between two nodes?
Me: (I knew the optimal solution (logn) for this, but I chose to tell him the normal solution which
was BFS.) Sir, BFS.
I3: Can you implement it?.
Me: I had used this algo quite a lot of times so it didn’t take me much time.
I3: Great.
I3: Moving on. Again, a question on tree. Draw a tree that has pairwise distance average equal
to 1.
Me: I thought for a few seconds and drew a tree with 2 nodes (not knowing that this is the only
answer)
All: Started laughing as they didn’t expect me to come up with this solution.
I3: Okay, forget this. Draw a tree with 5 vertices.
Me: I started drawing trees randomly and calculating the pairwise distance average. But this
was time-consuming, so after drawing 2-3 trees, I proved mathematically that any other tree is
not possible. (this went on for almost 5 mins)
I3: Good. Now, draw a tree such that the pairwise distance average is <2 for n=10.
Me: I made the same mistake again, I started drawing trees randomly and calculating the
pairwise distance average. But after 1-2 trees, I was able to draw a tree that was correct.
(Answer: one node is the root and all the other nodes its children)
I3: What observation do you make from this tree?
Me: (It took me a while to come up with the correct answer and they seemed satisfied with it.)
(Answer: The maximum distance is 2 and the minimum is 1, so the average will always be <2)
I3: Was really happy with this.
I2: Asked me questions on P&C and functions. (this prof tried to see how I think)
Me: I answered most of them orally but answered one question incorrectly, so I used the board
and wrote the solution.
I2: (happy to see how I arrived at the solution). Great, nice.
I1: Draw the graph of normal distribution.
Me: Drew it instantaneously.
I1: What does this function look like?
Me: I wrote the function on the board, again instantaneously. (As I mentioned earlier, it was a
wise decision to study probability)
I1: Great. Write down the Bayes Theorem on the board.
Me: I started by explaining them about conditional probability and tried to get the exact formula
from it. It took me a while and the interviewers were confused as to what I was trying to do.
I1: Why is it taking so long?
Me: Just a minute.
I1: *Throws 2-3 more questions related to Bayes theorem*
Me: Answered them orally and instantaneously.
I1: Correct, so why are you taking so long to get the formula?
Me: I didn’t say anything (I don't know why) and continued to solve.
(after 2-3 mins, got the formula)
I1: Correct. I don’t understand what took you so long.
Me: I was trying to find the expanded form of the formula and not the basic one (conditional
probability)
I1: Okay :/
I2: Asked a question on Isomorphism (group theory). (I don’t remember the exact question)
Me: Answered it correctly.
I2: Given two sets A and B of size n, count the no. of bijective functions.
Me: I knew the answer already but I solved it on the board, drawing functions and arrow and
stuff like that and arrived at the solution.
I2: Good (when I was drawing those things. again, wise decision)
I4: I asked a few questions on counting.
Me: Answered them correctly
I1: You can take your documents and leave.
Me: Thank You.
I3: Thank You (what?)
The interview went on for around 25-30 minutes.
Interview Experience 3:

In the whole interview process, I was the second last person to be called in. 4 panelists seemed
too tired to take my interview.
* I handed over my certificate folder to I1 and comfortably got into my seat*
I1: *Seeing my previous college name” IIITDM Kancheepuram” smiled*
Me: *Wondering what's wrong*
I3: Why are your college students very much interested in joining IIITH? (there were 5 students
for the interview from my college. Me last among all)
Me: There were few seniors from previous batches who joined IIITH, set trend in college
(smiling).
I3: *some more questions in specific to my college*
Me: *answered them*
I2: Which mathematics courses do you like?
Me: Discrete Mathematics, Probability. (I prepared well for these topics)
I1: Ok now take the marker and move to the board. (*Asked to draw a graph, Verbally giving the
connections between nodes*)
Me: I was slowly doing, Meanwhile trying to guess what the question was. (I understood that the
graph is a tree)
I1: What is the chromatic number of this graph.
Me: I immediately answered 2. And intentionally detailed my explanation saying every alternate
level of the graph has the same colour.
I1: Yes, correct.
I2: You mentioned graph theory in your interests, Tell me what is a spanning tree?
Me: (*I was confident in this topic and answered*) Spanning tree is a subgraph that covers all
the edges with minimum no.of edges.
I4: (*without any gap*)How many subgraphs does a graph have?
Me: I don’t know the formula. But I can give an algorithm to find it.
I4: That’s fine you can proceed.
Me: I was explaining and trying to be accurate in my calculations
(they understood that I could do it and said to stop me in b/w)
I2: Draw a complete graph of 4 vertices.
Me: I did it.
I2: Good, now Find the subgraphs of this according to your algorithm.
Me: I was comfortable doing the question.
(they continued questioning on graph theory)
(I2&I1): (* some questions on Degree Sequence, Isomorphism, etc..)
Me: I answered many of them correctly.
I4: Let's now get into set theory, give a proof for finding no.of subsets of a set.
Me: I answered them. ( And was explaining them using concepts of combinations, where I
messed up relating a set to function)
(Before I correct *I3 interrupted*)
I3: What is a bijective function? ( being little sharp)
Me: (Little tensed, But knew the answer) Answered and gave the formula for no.of bijective
functions.
(I3 continues....)
I3: You did a course in Linear algebra right? ( Seeing my grade sheets)
Me: (I don’t want them to question me on that) Yes Sir, But it wasn’t one of the courses which I
liked or were taught Properly.
I3: But you have a very good grade in this? (Taking ‘A’ grade to be top grade)
Me: No sir, I got a second grade. In my previous institute, Top grade was ‘S’ followed with A & B.
I3: (understood that I am trying to skip the question) Ok, But I will only question the basics. Are
you comfortable with that?
Me: I had to say Yes.
I3: (*questioned on some basic concept*)
Me: *I managed to answer correctly*.
I2: Stated a question related to conditional probability and asked to find the solution. (observing
what concepts will I use to reach the answer).
Me: (I did the same kind of question before) I Directly skipped to the final step giving the formula
to get the answer.
I2: Can you give the procedure for this?
Me: (I forgot how to get to the formula). Started writing some obvious things on board and trying
to find a clue.
(Fortunately *I3 interrupted again*)
I3: Draw the normal distribution curve.
Me: I drew the bell curve.
I3: What is the probability density of normal distribution?
Me: I don’t exactly remember the equation.
I3: (*two other questions which I don’t know*)
Me: *Two continuous replies* I Don’t know:/ (Losing hopes)
I1: You mentioned that probability interests you? And you don’t know many topics?
Me: Yes I got interested in it. But I am not exposed to it fully. (With trembling voice)
I4: *interrupted* Don’t worry, you are doing good so far. Can you try giving your thoughts on the
previous question? (this relieved me a bit)
Me: Then I was trying to explain all the guesses I had.
(They were ok with that).
I1: If you are interested in something, Then you should put your effort into exploring it.
Me: Yes sir, I am trying. I had also applied and got selected for a summer research fellowship
program in topic closely related to probability *Showed them the selection certificate*
(they looked at it and passed in on to other)
I3: Good, You can leave now.
Me: Thank you.
This all went around for 20 min.
Interview Experience 4:
Me (with a huge grin): Good morning!
They give an approving nod and I hand over my file with all my certificates. I2 - I4 go through
the file while I1 talks to me
I1: What is your name?
Me: Bharathi Ramana Joshi
I1: Where are you from?
Me: Keshav Memorial Institute of Technology, it's a college affiliated to Jawaharlal Nehru
Technological University -
I1: Yes, but where is that college
Me: Narayanguda
I1: Oh, so you're Hyderabad only
Me: Yes
I1: What are you interested in Computer Science?
Me: I'm interested in Computer Science research, specifically Programming Language Theory
and I'd like to work under Dr. Venkatesh Choppella and Dr. Suresh Purini (note to the reader:
they are both professors at IIIT Hyderabad)
I1 directs me to the whiteboard and starts asking questions
I1: Write a fragment of C code that swaps the contents of two variables
Me (writing while explaining): Well, there are three ways to do this - first we can use a third
temporary variable, second we can a series of arithmetic operations and third we can use
pointer arithmetic
I1: Is it more efficient to swap contents of the two variables, or the pointers to their addresses?
Me: It depends on the hardware implementation
I1: How so?
Me: If the variable size is greater than the pointer size, swapping pointers is more efficient.
Otherwise swapping variable contents is more efficient.
I1: Hmm. Now write some C code to transpose a matrix using pointer arithmetic only
I struggle to solve this for around ~5 minutes
I1: It's alright, try it after you go home. Now, do you know what Baye's theorem is?
Me: Yes, it is used to express conditional probability
I1: Good, write it on the board
I write down Baye's theorem, however I use a slightly different notation from what is given in the
textbooks
I1 (giggling, perhaps at my new notation): Where did you learn this?
Me: Intermediate textbook
I1: If two coins are tossed and you know that one of them is tails, what is the probability that the
other is also tails
Me: 1/3
I1: How can you say that? Work it out using Baye's theorem on the whiteboard.
I proceed to show my work on the board
I1: Good, now what is the probability that not both are tails?
Me: 1/2
I1: Please solve it on the whiteboard
I again show my working.
Now I1 starts to see my file while I4 asks me questions.
I4: Say there are six trees in a straight line, with one bird on each tree. You are allowed to move
a bird from tree-a to tree-b in one direction but you should also move another bird in the
opposite direction from tree-c to tree-d such that the distance between trees a and b and trees c
and d is the same. After some moves, is it possible to have all the birds on the same tree?
(Note to the reader, to clarify the question, if you move a bird from, say, tree-1 to tree-2, you
should also move a bird from tree-6 to tree-5 (or any other pair of trees separated by the same
distance))
I try solving it for ~5 minutes
I1 notices me struggling and gives a hint
I1: Try solving it for odd number of trees, say 5
I solve this very quickly and realize the question is unsolvable for even number of trees
Me: All the birds cannot be moved onto a single tree in case of even number of trees as after
some moves, birds keep moving between the same two adjacent trees.
I4: Good, do you know what the loop invariant here is?
Me: The distance moved
I4: Good
I3: How good are you at Operating Systems?
Me: I'd say I'm decent, given that my Google Summer of Code project comes under Operating
Systems
Interview panel visibly impressed, perhaps due to me mentioning Google Summer of Code
I3: Alright then, let's talk about hard disks. Can you explain how data is organized on a hard
disk?
Me: Yes, Operating Systems make use of b-trees to organize data efficiently
I3: Could you explain what b-trees are?
I draw a simple b-tree on the whiteboard and explain
I3: What is a b plus tree?
Me: b plus trees have extra leaf linkage to increase traversal efficiency, which b-trees do not.
I1: Where did you do your intermediate?
Me: Ramaiah IIT Study Circle
Interview panel visibly impressed
I1: Would anybody else like to ask any questions?
I2 - I4 nod a no
I1: You may leave now

And thus, my 26 minute interview (we had to sign in a log book while going in and coming out,
that's how I know :P) ended.
From Students Selected in 2019 (CSD)

Interview Experience 5:

Interview room was a conference room with one side 4-5 professors sitting and other side
was a seat for the candidate and a white board. As soon as I entered, they took the file of
the certificates and college results from me. I was first asked where I was from and my
college.

I was then asked for my favorite subjects to which I replied ML and its implication on
Distributed systems (which is also a broadly specifies my research topic)

They said to give core subjects to which I replied DS and Algo.

They then asked me to go to the white board.

Q: Suppose you have an array of N elements. You need to find for how many i, Ai + A(i+1)
is a square number. Is this question trivial?

Ans: Yes sir

Q: Now can you do it for any two numbers in the array?

Ans: Yes

Q: What is complexity of factorial

Ans: Exponential

Q: Can you reduce is to a polynomial

Ans: No

Q: Now suppose, In the array, all elements occur once except one which is repeated. Can
you find the repeated number:

Ans: Could you specify how big is the range of numbers?

Q: Specify for any range

Ans: For a small range, I'll make an array and in the array count the number of occurrences.
The element with >1 occurrences will be found in this manner For Large range, I'll go with a
hash table

Q: Tell that hash function you'll be taking

Ans: Any modulo hash

Q: But what about a....(some hand actions to which I implied collisions)

Ans: For collisions, I'll be doing chaining, and will go through it.

Q: But still going through that will be painful

Ans: Yes... I agree.... Let me think (scrambling on white board to come up with a bad algo
which I ended up not telling to them. Just started telling them like we'll be making them
random hoping that it is not sorted or reverse sorted and taking two pointers ... but then
stopped telling stating to them that anyways it is a bad method ) [ So now they kept
throwing some random questions when I was working out with problems on the white board]

Q: Ok so what are you doing currently?

Ans: I'm currently Interning at IIT Indore under professor <Prof Name> Sir.

Q: Ok leave the previous question.[I stop writing on the board and face towards them] What
are you doing there.

Ans: Working on diffusion techniques for learning over network.

Q: What type of networks?

Ans: [ confused] network.... with agents? [ At this point one of them was visibly upset by the
performance till now]

Q: Ok leave it. So you said your favorite subject is ML, so what have you studied in it?

Ans: Regression, Classifier, perceptron, SVM [ I was getting a bit unconfident now and thus
kept constantly blabbering and forgetting to specify many things or to say some random
response without thinking much]

Q: Have you studied random forest?

Ans: No [ Visibly upset professor is even more upset and is like why do they even have me
there]

Q: Okay so what types of regressions have you studied? [ BTW, these questions were
constantly being thrown from 3 of the four professors while one was just going through the
papers I gave him (at least this is what I think was happening]

Ans: Linear and logistic

Q: Tell me about Linear regression

Ans: Told them [with some errors which I later corrected (I initially forgot to augment 1 in the
feature vector )]
Q: Give me a closed form solution to this instead of iteration

Ans: We'll take mu as inverse of Hessian Matrix [ Initially I responded by just saying
"Newton Descent will be used instead of Gradient Descent" but after some hints understood
what they were asking]

Q: Ok do you know what rank is?

Ans: Number of non zero rows in echelon form of matrix. [I should have given them the
visual and logical answer if my mind hadn't gone black]

Q: for matrix: 1 1 1 2 2 2 0 0 0 What is the rank

Ans: 1

Q: reason?

Ans: [ explained]

Q: make it a 2 rank

Ans: [ edited last 2 to 3]

Q: give and example of 0 rank matrix ["example" well I guess they meant the size could
vary]

Ans: wrote a zero matrix

Q[upset guy]: for a rectangular matrix of NxM what is the maximum rank?

Ans: N

Q: even for M<N?

Ans: [ a bit of thinking ] min(M,N)

Q:Why?

Ans: [ making hand actions ] because in the other form of echelon, [ random blabbering] (I
must have explained this also logically)

Q: Suppose for a column vector A, Nx1, what is the rank? Ans: N [ no it is not]

Q: [ trying to give hint] what is there is a zero? Ans: let me check my answer if it is correct [
writes on whiteboard] N - Number of zeros [ no it is not]

Q: ok go... What is your CGPA?

Ans: 9.65
Overall: My interview was way too bad. BTW, selected for LEEE CSD

Interview Experience 6:

My Interview was the first one after lunch. There were probably 8-10 people more after
me. I entered a room which had 2 profs. I had my file in hand and sat down.

Faculty A : Tell us about yourself.

<This is the opportunity to put forth any accomplishments, achievements, your interests,
what you're passionate about and anything that can convince them you're inclined
towards research>

But I totally messed it up, I regretted it so bad afterward. I just said, I'm doing CSE in
Mahindra École Centrale and paused. I shouldn't have PAUSED! I should gone on and
on about fields I'm passionate about in Comp Sci, like NLP and DL and my previous
internships and MOOCs I was doing, anything. But no. My brain didnt work. And that
pause made sir ask his next question

"What is your favourite topic among your 4 semesters at college so far?"

Another very important thing is your temperament before and in the interview-room.
Keep yourself calm, because once it's over, it's over. We can't change anything. The
mind might go blank and simple things may seem incomprehensible, so stay calm, keep
it real and don't try to be oversmart, or put up a facade during the interview. It always
fails, they can tell from your answers.

My interview lasted for less than 20 minutes. I said my favourite topic is "Discrete
Maths" and was asked questions by the faculty

From Students Selected in 2020 (CSD)

Note: The interview experiences of students selected in the years 2018 and 2019 was
before the COVID-19 pandemic, and hence was conducted at the IIIT-H Campus. However,
candidates who qualified the written examination in 2020, were interviewed online, over a
video call.
Interview Experience 7:
Due to COVID, the interviews this year were online and followed a slightly different pattern than
last year(s). To prepare for the interview I followed the usual book recommendations for Discrete
Mathematics and Algorithms.
Two interviewers were connected but only one spoke

Interviewer: I, Me: M

I: Can you please turn on your camera?

M: Sure. Give me a sec

I: What topics do you like?

M: I am currently exploring Machine Learning, more specifically Deep Learning. I have delved
into its applications in visual tasks, auditory, <Interrupted>

I: Let me clarify a bit, I am asking you what topics would you like me to question you on?
M: Ah. Well, I have a fair amount of experience playing with Computer Vision applications that
use Deep Learning, etc.

I: Can you define non-linearity?

M: In the context of deep learning, we apply non linear activation functions in the intermediate
layers.

I: Why?
M: If we don’t do that, all the linear functions can be collapsed into a single function and we will
be wasting a lot of parameters.

I: Great. Now can you tell me what is an activation function?

M: Well…it's a function.

I: Anything else you would like to add?

M: I am not sure what I can add. Activation functions are used in neural nets, intermediate
layers mostly use ReLU these days but tanh was also commonly used. For classification tasks,
the last layer uses sigmoid.

I: Why sigmoid?
M: It normalizes the output as a probability distribution.

I: What do you mean normalize?

M: It transforms the output such that everything sums to one and individually lie between 0 and
1
I: But what does normalize mean?
M: I am not sure I understand your question.

I: Leave it. Can you draw sigmoid for me?

M: Sure. <Draws on paper and holds it up to the camera>

I: It’s wrong
M: Hmmm. Let me check.

I: What’s the range of sigmoid function?

M: 0 to 1?

I: Check the range in your figure.

M: Ohhhh. Yeah <Draws and shows a new figure>

I: That’s correct. Can you draw the log function?

M: <Draws and holds the paper>

I: This is log?
M: Yes?

I: Oh. Yeah. It’s correct

M: :-)

I: Do you know about Siamese Neural Networks?

M: I have heard that term but never really explored that topic

I: No problem. Do you know what a CNN is?

M: Yes

I: Great. Why is the advantage of using a CNN compared to a Fully Connected Neural Network?
M: Well the design of a CNN is much better suited for spatial data like images and audio. It
properly utilizes the fact each point is closely related to points near it than those far
<Interrupted>

I: Yes but you are missing the key reason

M: <After thinking for a second>. Less parameters?

I: Exactly!
M: Yeah

I: Do you know what is a black box algorithm

M: In terms of ML?
I: Yes
M: We consider an algorithm to be black box if we cannot tell what caused it to produce a
specific output. For example CNN are considered to be black boxes. I somewhat disagree with
that sentiment though. There has been a lot of research in understanding CNNs and we can
apply a lot of techniques to visualize what parts of an image a CNN is considering for an output.

I: But even with these techniques CNN does not have the explainability of white box algorithms
M: That’s true I guess

I: Ok. Now can you tell me the formula for Cross Entropy Loss?
M: Let me try. <After a minute of struggling>

I: Can’t recall?
M: I don’t remember the formula because I don’t have to code it manually a lot. Anytime I have
to use it I just look it up. Most of the time, I use the predefined one in libraries like PyTorch and
Keras.

I: It's alright. Tell me why do you want to join IIIT

M: IIIT is one of the best research institutes of the country and I would love an opportunity to
work and learn under some of the leading researchers in the field.

I: Great. That will be all

M: That’s it?

I: Yeah
M: Thanks for your time

Some remarks:
I spent the two weeks from result declaration to the interview studying mostly discrete
mathematics because it was my weakest and from reading the past interview transcripts it
seemed like most questions will be based on that.
I was caught off guard when he told me to pick the topic for the interview and I just picked what I
thought I knew enough about.
I was disappointed because I couldn’t answer some of the basic questions and didn’t really have
much hope but was pleasantly surprised when the results came out.

Interview Experience 8:

Exam prep :
I studied only for a week that too 4 weeks before exam date ,since I was occupied with my end
sem assignments and exams.In Fact i had my end online exams just a day before and after the
lateral entry exam,but i just went for it.
I have revised discrete mathematics , data structures and algorithms ,focusing more on
graph theory (source : lecture notes of previous institute ) but unfortunately I didn't study some
standard algorithms which were not taught due to lockdown but they appeared in the
exam(bellman ford and floyd warshall).
I revised combinational circuits from digital logic( source :gfg and neso academy) ,I
didn't revise sequential circuits enough ,but luckily there were no questions from that part
.Overall, the Digital logic part in the exam was easy.
A mistake which i made and advise you guys not to do is ,don't neglect c programming
,this is the part i felt trouble for ,i was very confused and made errors.i suggest everyone to go
through pointers , function calls ,recursive calls, passing of pointers from one function to another
,and what happens when you use same variables for passing and then you modify it and all.

Computer organization was taught in our college using armsim ,but generally it would be
on x86 processor, so i didn't revise much as everything would be completely new ,i just revised
basics .As far as i remember there were 3 questions and 1 was from basic ieee 32 bit
representation and 2 questions (obviously) x86 processor related which i randomly marked
some option.

Exam:
I messed up c programming a lot ,but did not waste time thinking about those.Instead i tried
solving aptitude problems which i was not able to solve in first go ,and to my surprise i have
solved around 8 new questions.I would say just do not panic if you don't get something ,try to
get maximum out of the paper ,be it whatever topic it is, since all questions have same
weightage.

Interview prep:
I was so scared .I didn't have internships /projects /achievements and all.I was not even
into competitive programming. .I didn't even have any online courses to mention . I was part of
hpc labs at my previous institute but didn't mention those as there wasn’t any good progress.i
didn't even make any resume.All i had was a decent cgpa of 9.09.
I wanted to choose discrete mathematics ,so i revised it all again .Along with it i revised
probability as a backup(after seeing some of previous year interview transcripts).But
yeah,something unexpected happened in the interview.

Interview
I got a link for the interview around 3pm and was admitted at 3:25pm.
At first their voice wasn't clear.When i told them ,they asked me to unplug headphones and
check, i told them it was even worse , they laughed(i had no clue why) and asked me to plug
headphones and continue..

Q:So ,where are you from?

Me: I live in mahabubnagar and my previous college was IIITDM Kancheepuram

Q:Oh! Do you know that your seniors are studying here?

Me: yes sir ,i came to know that after clearing the exam.

Q:Nice,so what do you do apart from academics?

Me: I play volleyball sir.
(just came out from my mouth without even thinking properly).

Q:So ,you are into sports that's nice(they were so impressed),how often?
Me: (i wasn't expecting that) I don't play if there is a lot of work or else i will play regularly.

Q:so , which topic do you like?

Me:Discrete maths , probability( in low voice)

Q:and ?
Me:Data structures but..(mumbled something I guess I wasn't audible to then so they
continued.)

(so i was scared as i didnt even revise it)

Q: So you know facebook right?,and people make friends with new people ,so you need to
store these connections in such a way that if you select a person all his friends should appear?
Me: maybe hash table ( this may seem simple now , but i was scared and i remembered that
array ,with open chaining structure and let it out as hashtable)

Q:No,don't say maybe ,think properly(they said something , i don't remember it)
Me : adjacency matrix.

Q:Yes, you should say graph data structure,ok then what are two ways of representing…
Me: (interrupted and said) Adjacency matrix and adjacency list.

Q.Yes,which is better?
Me:Adjacency list.
Q: why?
Me: much memory will be wasted if you use adjacency matrix.
( we had some discussion regarding this ,they seemed fine with it.)

Q:(definitely other people and audio quality became very bad) so consider a complete graph ,
do you know what a fully connected subgraph of a graph is called?
Me: yes sir, clique.

Q:so you have a completely connected graph ,

1) you start with a vertex ,and come back to it only when you travelled all other vertices.
2) You can have a sub traversal of traversal repeating but not the whole traversal.
How many such traversals are possible?
( ps: i am not sure if this was exact question, i already asked him to repeat it 3 times due
to bad audio ,which became even more confusing , and sir said at last “i complicated it
i guess , you can simplify and say the answer”)
Me: ( I asked if it's undirected and they said yes,after some work I gave the answer as n! he
said there would be more ,later he said ok we will move on to the next question.)

Q:name two real world problems that you want to solve.

Me: ( i was tensed so much and my mind went completely blank, took a min to answer )
I said something like performing fast and efficient computing at the stock exchange .

Q:Name 3 subjects that would be needed to do that.

Me:Operating systems-to know how tasks are scheduled by os, computer architecture-to know
how processors work at low level.
( they asked if they both are the same , to which I replied they are two different courses in the
institute so I thought they would be different, and i was completely blank and dumb now , i was
only thinking of courses that weren’t completed)
Computer networks - to efficiently store and process data ( i should have said database
management or algorithms or atleast should have given proper reasoning)
Q:computer networks! Hardware or software?
me:(slowly) software.

Q:Then ..isn't computer networks hardware?

Me: (mumbled) high performance computing

Q:ok ,you can go ( and disconnect the call).

It felt like it was completed in 10-15 mins. I thought the interview didn't go well , but I
guess you can't conclude anything till results.

Interview Experience 9:

Q: Tell us about yourself. What are your hobbies?

Me: Hello. I like to participate in contests involving competitive programming and machine
learning.
I told about my friends, sports, etc

Q : Okay. That is about your last two years. What were your hobbies before?
Me: I like to draw. I have been drawing since childhood. I have a drawing with me. Should I
show it?

Q : Yes show it.

It was a pencil sketch of Genos. One of them recognized it to be from an anime.

Q : Why are you doing engineering? Why did you not join an art college?
Me : I was startled. I spoke whatever I could think.
There is a lot more to learn in computer science…

Q: I have N nodes which are pairwise connected. How many possible paths are there of length
k?
Me: N^k

Q : Now, each of these edges have a cost. Suppose total reward is the sum of these costs. How
to find a path with maximal reward?
I thought of the Djikstra’s algorithm. Maybe if I set the cost as negative, it will give the maximum
cost path instead. But then I realized that it won’t work for negative weights. Bellman Ford could
be used maybe.
Me : We can search for a positive cycle. We can keep increasing our cost by going through the
cycle repeatedly.

Q: How to reverse a singly linked list ?

Me: Take three nodes, say prev, cur and next ...You can find the answer on geeksforgeeks.
Q: Do e^x and ln (x) intersect?

Q: How to reverse a doubly linked list?

Me : Why would we require it? We would (almost) get the same list.

Q: How to define dynamic programming?

Me: Like to find fibonacci’s numbers we can use dynamic programming. We memoize the
result…
If a problem has optimal substructure we can use it.I was not able to solve it then.

Interview Experience 10:

For the exam, I revised discrete math and algorithms fairly thoroughly, just used standard
textbooks like Rosen and CLRS. Was already comfortable with C, didn't do anything special for
that. Did solve GATE questions for every topic.
For the other subjects, digital logic and computer organization, I just relied on solving previous
GATE questions. Didn't bother with the theory much, except for topics which I felt were
commonly asked about. Used online stuff like NPTEL for that. Skipped difficult questions
because I didn't think it was worth the time investment.
I didn’t do any preparation for logical reasoning.

Had been studying on and off for the better part of the second year, but really only sincerely
practiced in the last month. That I didn't have college to worry about courtesy of the lockdown
helped. But I was reasonably familiar with most topics to begin with by just following college
courses, so most of my preparation was just revision.

To have some concrete topics to talk about in the interview, I revised linear algebra, probability,
and algorithms. Of these, only linear algebra came up.
As for the achievements and experience, I had qualified for ACM ICPC regionals, had
completed the deep learning specialization course on Coursera and had worked on 3-4 machine
learning projects, some individually for learning purposes and some for coursework (weren't
anything special, just padded the resume). I was also interning under a professor at that time for
an NLP-based project. Listed all of these in my resume. Only a couple of these came up.
Transcript -
Interviewer 1: We have someone coming in from the NLP lab, we will wait for them before
beginning. In the meantime tell me something about yourself.
Me: [Introduce myself and basically mentioned every topic I am interested in]
I1: What are your hobbies?
M: [Mention some generic hobbies. The next few questions were about the mentioned hobbies.
They were just questions to help me get settled. I had also submitted my ACM-ICPC regional
certificate, he asked about my performance in it]
At this point the person from the NLP lab had arrived.
I2: So you’ve mentioned that you’re working on a project under [Professor]. Can you describe
your work?
M: [Describe the work, they inquired a bit about how I got the opportunity and asked general
questions about the topic]
I2: What are embeddings?
M: [Give my understanding of embeddings, say a few dumb things because of nerves which
they ask me to clarify]
I2: You’re only talking about word embeddings, what do embeddings mean in general?
M: [Say that I am familiar with only word embeddings, but try to give a general answer as well]
I2: You’re closer, but it’s still not right. Describe a few properties of embeddings. You’ve already
stated one, that they’re numerical.
M: [Was completely lost on this. I thought of a few things, but then just said that I don’t know.
They prompted me by giving some examples, like dense or sparse, whether they should be
discriminatory in nature. Elaborated on those two headings but couldn’t think of any on my own]
I2: You should be able to think in more abstract ways. Try to find more later.
I1: Since you qualified for ACM-ICPC, you must be familiar with data structures and algorithms,
so we won’t ask any questions from that. Are you familiar with object oriented programming?
M: [Said that I am familiar with Java to a certain extent and described object oriented
programming briefly. I should have said Python since I am far more experienced in that but for
whatever reason brain just completely blanked on the fact that Python is also object oriented]
I1: Since you say you’re familiar with Java, can you describe how to declare objects without
using the new operator?
M: [Said I don’t remember such specific details, since I worked with Java a few semesters ago
for an OOP course]
I1: Do you know what a clone is in Java?
M: [Explain what clone is, basically just an identical copy of an object. But said I’m not sure how
to use them for declaring new objects]
I1: There’s many ways in which new is implicitly used. Search for those later. Do you know what
virtual functions are?
M: [Describe them in context of Java]
I1: Are there equivalents to virtual functions in other languages like C?
M: I am only familiar with virtual functions in object oriented languages, I’m not sure what it
would mean in context of a language like C
I2: What is a pointer function?
M: [Explained what it is, understood what they were hinting at, and explained how pointer
functions can be thought of as virtual functions]
I1: What do you know about functional programming languages?
M: [Describe them, struggle to come up with an example, but they help me with that. Also clearly
state that I have no experience working with functional languages]
I1: I’m going to make a statement now, tell me whether you agree or not - anything that can be
done in object oriented languages, can be done in functional languages.
M: It isn’t water-tight reasoning, but the fact that languages like C exist which aren’t object
oriented tells me that that is true. Object oriented programming is just one paradigm, but isn’t
necessary for any task.
I1: Which one do you prefer, object oriented or non-object oriented?
M: [Explain my preference for object oriented]
I2: Okay so let’s move on from that. What is a basis function?
M: [Wasn’t sure what this meant] Could you please repeat that?
I2: A basis, like a basis vector. Can you define them?
M: [A basis vector I knew. Defined it but I did it in the context of a matrix. They then asked me
define it for a vector space, not for a matrix]
I2: What’s the condition that basis vectors follow?
M: [Mention linear independence. Also give their relation to the dimension of the vector space
for good measure]
I2: What if they’re not linearly independent?
M: In a matrix, we will get a row of zeros in the row echelon form.
I2: You’re talking about matrices again. What does it mean in general?
M: [Explain that if they’re linearly dependent vectors then some vectors won’t give any new
information and are redundant]
I2: What’s the condition for linear independence?
M: [Give the mathematical condition. They ask what happens if there’s exactly one non-zero
multiplying constant. Don’t think I gave the right answer to this]
I1: Okay, one last question before we end. People often say that Java doesn’t have multiple
inheritance. Why is it so? Does Java support multiple inheritance?
M: If I recall correctly, it doesn’t explicitly support multiple inheritance but the effect can be
replicated using interfaces.
I1: Fair enough.

The interview ended after that. It was 17 minutes long.

Didn’t expect that ACM-ICPC would allow me to completely bypass DS and Algo questions, I
had been under the impression that these would be important topics for the interview.
I was a bit disgruntled by the technical questions on Java, even after I had said that I am not
overly familiar with it. But in hindsight I shouldn’t have mentioned Java at all if I wasn’t
comfortable with it.
Still, while the questions asked did take me a bit by surprise, they were all from topics that I
knew to some degree through just college courses. I struggled with quite a few questions but
the interviewers kept trying to nudge me towards the right answer. Their tone was amicable
throughout the interview. I asked them for clarifications wherever I felt was necessary. They
were probably willing to overlook some inaccuracies in the answers if your thought process was
right.
But the fact that it wasn’t really possible to ask questions requiring written work made it different
from previous interviews. I was asked purely theoretical questions; writing algorithms or solving
numerical questions on the spot would’ve been a bit trickier. If you’re living in a post COVID
world and are going to have an offline interview, then your experience will probably vary
significantly.

Interview Experience 11:

Hey. Below is my interview transcript (LE - CSD)

My interview was I guess one of the first one’s, it was in the morning at 9. It was an online
interview, thanks to corona.
Personally, as preparation, there really wasn’t anything in particular that I had done. Just a day
or two before, I just had a mock interview with an Uncle I knew.
So I joined the meet and I was waiting to be admitted in.
There were three professors, let’s call them P1, P2, P3. (my memory is a bit vague as to who
asked what, so the order could be a bit messed)

Me: Good morning Sirs.

P1: Good Morning, [and asked me to introduce Myself and why I want to join IIIT]
Me: [Told about myself and my previous college, ans about why i looked forward to joining in
IIIT, what I liked there why i liked research And something on these lines]
P1(or P2): That’s Good, Now tell me what subject/Topic are you interested in.
Me: Sir I like Data Structures and Discrete Mathematics.
P1: Let’s go with Data structures. Which one do you like?
Me: Graphs sir
P1: Think and answer, you’re going to be questioned on that.
Me: Yes sir, I am sure.
P1: Ok tell me the different types you could represent a graph.
Me: Adjacency Matrix, Adjacency List, and an Adjacency Set.
P1: Which one do you like.
Me: I Like Adjacency List the most as we have dynamically allocated space for edges and [
something here ]
P1: Tell me about the directed and undirected graph
Me: Told him.
P1: Take an example graph with three nodes, draw it on paper represent both in adjacency list
form and show me.
Me: Drew the graphs and pointed it to the camera.
P1: Which algorithm do you like the most in graphs
Me: Sir I kinda can’t pronounce it that well it kinda goes like Djkastra (I said with a weird
pronunciation)
P1: (Laughing) Yeah I know it’s confusing I call it (... Told how he pronounces), anyway why do
you like it.
Me: Sir I was recently doing a course in Advanced Data structures on Java that covered graphs
and algorithms, and out of all I liked Djkastra the most as it gives us the shortest path while
keeping track of kinda all other paths, and if you keep distance equal for all it ends up becoming
Breadth First search and that got me interested and intrigued about it.
P1: Good Good, that’s it from Me.
P2: Hey, so you did say you like a bit of Discrete, which parts of it?
Me: I like PnC and probability.
P2: Asks me a PnC question (Something on choosing and it’s probability)
Me: Answer it.
P2: What is an injective function ?
Me: I don’t exactly remember the terms of relations and functions but I guess it goes like
(Answered something wrong)
P2: Ok draw a graph of xe^(-x).
Me: Draws it on paper and shows him.
P2: Asks a simple probability question.
Me: Answer it.
P2: Good, That’s about it.
P1/P3: Is there anything you want to tell us about.
Me: I talk about a kind of project I work on (Custom ROM’s)
Them: Ask me what I had contributed and what bugs I found and all.
Me: Tell them about one I found in their official release and the steps how I fixed it and all.
Them: Ok that’s it. You can Leave now.
Me: Thank You Sirs.

(It was over in less than 20 mins, IT didn’t seem bad, but also I thought it wasn’t good enough
either. But a couple of weeks and I get a mail that I got selected).
That’s how my Interview went.

Interview Experience # 12:

As I read interview transcripts for seniors, I found that most of them were asked about their
favorite subject and then were asked questions related to that. So, I started revising formal
languages and automata theory during the time between the declaration of exam results and the
interview.
During the interview, firstly I was asked to introduce myself and to tell me why do I want to join
this course. You should always be prepared for such questions because they are frequently
asked. I started by telling my name, background, interests, what my previous college was
lacking, and what IIIT-H provide me. Then as to what I predicted they asked me about my
favorite subject. Most of the questions were direct like what is a regular expression, regular
language, Chomsky hierarchy and some required a bit of thinking. This part of the interview
went well for me.

Then they asked me what all maths subject I had. I told them calculus, linear algebra. Though I
shouldn’t have mentioned calculus because I had it in the first semester and I didn’t remember
things well. They asked me to draw a rough graph of the equation given. I wasn’t able to draw
the complete graph but I told about some properties that the graph will have based on the
equation. Then they started asking about linear algebra (linear independence, basis, etc).

Then they told me to choose a subject from the set. I chose Computer Architecture. They asked
me to implement OR gate using a multiplexer and why a look-ahead carry adder is better than a
ripple-carry adder.

My interview was about 30 minutes.

A tip that I want to give to all aspirants is that confidence is the key. Think well before saying
anything because that thing can be their next question and prepare a list of subjects in which
you are comfortable.

Life at IIIT-H (Academics)

Q: How is academics at IIIT, is it very difficult?

A: If you’re coming from a college with an extremely laid back culture, you will find the
academics at IIIT-H quite gruelling. It is not exactly difficult but you should be prepared for a lot
of work, and a lot of programming.

Q: Will I be able to catch up with the other people who have been studying since the first year?
A: IIITH introduces several core subjects in the first year, while most other colleges tend to have
science and engineering courses in the first year. There will be a few subjects that your peers
would know and you wouldn’t. But don’t let that deter you. IIITH allows you to audit (attend the
classes of) any course, so if you’re willing to put in the extra effort, catching up won’t be an
issue.
Q: How does the research program work?
A: The student chooses a research advisor at the end of their 2nd year. The student works with
the advisor during the period of study. Over this period, a student must successfully complete,
along with the courses, 24 Thesis credits. In order to achieve this, advisors generally expect the
students to stay on the campus during the summers and do research. Keep in mind that
graduation depends solely on when you complete your research, which may take more than a
minimum of 4 years.

Personal views (Note: These are the views of the authors of the doc and does not reflect
general opinion)
Q: Is it worth joining, even though I will join in 2nd year?
A: It completely depends on what you want to achieve by joining IIITH. I would generally advise
against joining just for the job prospects since the graduation time is variable and the workload
is quite intense. It’s much easier to build up a resume through activities such as projects,
open-source or competitive programming and apply for jobs off-campus than go through IIITH’s
intense curriculum.
However, if you have a genuine interest in learning or research, especially in areas like
computer vision, robotics, and AI, then it is definitely worth joining. The exposure you will get to
research at the undergrad level is unparalleled in India.

Q: What’s the difference between this and dropping a year to give JEE/UGEE again?
A: The syllabus of the LEEE is the same as what you would typically cover in a CS program, if
you are comfortable with that you should be able to crack the exam without much extra effort.
JEE/UGEE, on the other hand, has a completely different syllabus, which needs extra work from
your side. It’s also worth noting that LEEE comes with significantly less risk than dropping a year
for JEE/UGEE since you have the backup of continuing your degree.