2) M.Tech in Computer Technology – Cutoff ( to apply): Gen/EWS-550, OBC- 522
3) M.Tech HVA CSE - Cutoff ( to apply): Gen/EWS-700, OBC-630
4) M.Tech HVA Computer Technology
5) MS-Research Machine Intelligence And Data Science ( MINDS,Yardi School of AI -ScAI)
6) MS-Research Amar Nath And Sashi Khola School of Information Technology (SIT)
ADMISSION MODE :
1) M.Tech in Cyber Security – Direct admissions based on Gate Score
3) M.Tech HVA CSE – A shortlisting in-person written test followed by interviews.
- I gave the online interviews for M.Tech Computer Technology HVA and MS-R SIT.
- For the MSR Yardi School of AI, I couldn't appear for the written test as I forgot to fill up the Google form ahead of the written test.
- For M.Tech HVA CSE, the written test was in-person, and I chose to appear for other in-person programs at other IITs that time, hence, I couldn't attend it.
- The dates for the written exam and interviews will clash for almost all the IITs/IISc. You will have to prioritize things, make a choice, and arrange your travel plans accordingly.
Me: Told briefly. I also mentioned my interest in the ML/AI domain and my final year project, which was also based on Machine Learning.
Interviewer: What subjects have you studied in your B.Tech?
Me: (Told the subject names I had prepared well for GATE and interviews) Operating Systems, Database Management, C programming, Engineering Mathematics, and COA.
Interviewer: What is the difference between a Process and a Program?
Me: A Process is a program that is in execution, whereas a Program is just some set of instructions or code stored on disk. A program that is running in RAM is called a Process.
Interviewer: What are NP-hard and NP-complete problems?
Me: Ummm..Sorry, sir, I don't remember this.
Interviewer: No worries, tell me why time complexity is measured asymptotically, and why it is called asymptotic?
Me: Time complexity of an algorithm is measured asymptotically to know how much maximum time it takes if we give n inputs to it, and at what minimum time it may finish. There comes the concept of Lower and Upper bounds. The Upper bound, also known as the Big Oh, tells us what maximum amount of time the algorithm may take for n inputs, whereas the Lower Bound, Big Omega, tells us at what minimum amount of time the algorithm may give output for any n inputs.
Interviewer: Do you know Machine Learning? As you mentioned, you are interested in the ML/AI domain.
Me: Yes, Sir, I know only the basics of Machine Learning, but I have a good grasp of Linear Algebra, which ultimately forms the basis of Machine Learning.
(The Professor was trying to delve into Machine Learning concepts, but I tried to mold the discussion towards Linear Algebra as I had prepared that well.)
Me: (He started asking Linear Algebra 😄) The Rank of a matrix tells us the number of Linearly Independent Columns or Rows in a Matrix. It gives the number of Linearly Independent vectors we have.
Interviewer: What do you mean by Linearly Independent Vectors?
Me: The vectors that cannot be written as a Linear Combination of other vectors in that set of vectors are called the Linearly Independent Vectors i.e if we are not able to find any non zero ci such that c1v1+c2v2+c3v3+...cnvn=0, then, the vectors v1,v2..vn will be linearly independent.
Me: All the Linear combinations of the columns of matrix A produces the column space of A.The number of linearly independent columns in the matrix A gives the rank of the matrix.
Interviewer: How to find eigenvalues for a non-square matrix?
Me: ( I thought about it for a while. I felt like it's a trick question, but I was not sure, so I thought a bit more. Actually, we can't find eigenvalues for a non-square matrix.)
He was satisfied with the answer and said, “Very nice, you’re good :P.” Then he asked, “We’re done with the questions, do you want to ask anything?”
I had faced many rejections before this interview, and thus, although it felt like the interview went good, I didn't keep any expectations!
Interview Insights :
Me: Good morning, Sir. I am Kriti Rathore. I've done my graduation in Computer Engineering from Bharati Vidyapeeth College of Engineering, Pune. I've done my final year project in Machine Learning, and I am inclined towards working in that domain.
Me: (this was not expected 😂 I took a moment to settle in. He said, "you can use pen and paper", so I started calculating 1/7 in decimals) Umm..0.14285.
Me: No, Sir, it goes on infinitely.
Me: (After thinking for some time, I got reminded of the repeating decimal concept that we studied in class 10th. So, I calculated more values after the decimal.)
Yes, sir. The values are repeating after some digits. So, this will be 142857 bar after the decimal.
Interviewer 1 : Will the digits repeat in all cases?
Me: No, Sir. They don't repeat in all cases. We have two types of decimal values, terminating and non-terminating repeating. ( I vaguely remembered this terminating, non-terminating thing from class 10th RD Sharma lol.)
Interviewer 1: If it's non-terminating, will it always repeat for all rational numbers?
Me: No, Sir. It won't always repeat. Like for example, if we take pi, i.e,. 22/7, it goes on infinitely like 3.14...and doesn't repeat.
Interviewer 1: Well, it is either terminating or non-terminating repeating for all rational numbers.
Interviewer 2: Also, pi is not equal to 22/7, it is an approximation to it, and 22/7 gives a non-terminating repeating decimal value.
Me: ( Fearfully) Okay, Sir 😐
( I messed it up here badly. I totally forgot that 22/7 is an approximation of pi and not pi itself.
Interviewer 1: Okay, how do you find out if a decimal is repeating?
Me: We keep dividing the numerator by the denominator till we get a pattern.
Interviewer 1: How do you find a pattern?
Me: Clueless*
Interviewer 1: Okay, take a number and divide it by 7, what remainders will you get?
Me : (I was quite nervous till this time, I had no clue where the interview was heading, so I tried to do what he said using pen and paper.)
I took a random number and started dividing it by 7 and then instantly said 0,1,2,3..till 6.
Interviewer 1: Okay, now divide a random number by 7.
Me: I continued dividing the number.
Interviewer 1: Do you see any pattern?
Me: Yes, Sir. The remainders started repeating a certain pattern after some time.
Interviewer 1: Yes, now can you write a program to find out if we get a repeating decimal or a terminating one?
Me : ( I went silent, I had zero coding knowledge)
Interviewer 1: Just give me an idea or algorithm of how it will work, you don't need to write the entire code.
Me : ( I tried to form a function that can detect the repeating pattern)
After some time, I said, We can use the modulus function to find out the remainder, and then check whether the remainder repeats the same pattern after some time.
Interviewer 1: How will you check that?
Me: We can use a for loop to find the remainders and then store them in an array, and check whether they repeat after some time.
I had no idea what was going on. I just said whatever I could think of at that time. I knew this was not going well 😕. I just told myself to hang in there.
Me: A linear function is something which satisfies f(x+y) = f(x) + f(y) and f(kx)=k.f(x).
A non-linear function does not satisfy these properties.
Interviewer 2: Is EXOR a linear function?
Me: ( I started thinking about it, but I went blank) No, Sir. I don't think so.
Interviewer 2: Are you sure? Take pen and paper and check.
Me: The exor function for A XOR B expands to A'B+ AB'. So, I don't think this expression is linear.
Interviewer 2: Okay, Kriti. This is wrong. Although you got the first part correct. Alright, you find it out by making a truth table and get back to me.
Me: Okay, Sir.
I had lost my entire confidence till now, and I was just somehow hanging up to what was going on, getting blank sometimes, like I did here. EXOR is a linear function. When we make its truth table, it is clearly visible that it satisfies the linearity property f(x+y)=f(x) + f(y).
This struck my mind immediately after the interview, and I felt very bad :(
So, these kinds of things happen during interviews, but you need to somehow be there and still tell yourself to focus on what is being asked and try to find the solution. The professors sometimes try to examine whether you are trying to think or just giving up at once. So, try to keep up with them, even if it is not going well.
Interviewer 2 left the meeting as he had to go to another panel to take interviews.
Interviewer 1: Okay, Kriti. So, as you have applied for MSR, where does your interest lie?
Me: ( With a sense of relief) Yes, Sir. I want to learn Machine Learning in more detail. This area excites me. I also made my final year project based on Machine Learning itself. So, I would be more than happy to explore this area in depth.
Interviewer 1: So, you know Machine Learning?
Me: Umm.. I mean, I have some knowledge of the basic models of ML, not in too depth, but I have tried to cover Linear Algebra in depth as it is the basis of ML.
Interviewer 1: Okay, explain what linear regression is?
Interviewer 1: Can you name some of the loss functions?
Me: (I totally forgot the names of the loss function) Umm.. No, Sir.
Different types of loss functions are Mean Absolute Error( MAE), Mean Squared Error( MSE)
Interviewer 1: What is a loss function?
Me: Loss function is a mathematical method used in machine learning to measure the deviation between the predicted values and actual values.
Interviewer 1: How to find it?
Me: We can calculate it by squaring the differences between predicted and actual values and then averaging them.
Interviewer 1: Okay, cool. We are done with the interview. We will get back to you.
3. For motivation on low confidence days and to keep a track of whether you are on the right track or not, Amit Khurana Sir's " Beauiful Tips for Gate Aspirants", "Honest Talks" videos are quite recommended. I myself used to follow these during my preparation.
7.Some journeys turn out to be the most beautiful chapters of our lives, regardless of the destination. So go ahead and give it your all. This could be the one for you!






congrats kriti! and nice blog!
ReplyDeletethanks for sharing this much in detail
ReplyDelete