Interview Experience ( There were 3 Professors, I'll address all three as P)
(Resume wasn't needed for Interview, so they didn't knew much about background of work that I've been doing)
P: Tell us about Yourself.
Me: Did.
P: How many questions did you do in Programming Test?
Me: I did 4.5
They seemed quite impressed by it (all three panelists started looking at each other)
P: What are your favorite subjects?
Me: DS, Algo, CN.
P: So should we start with DS?
Me: Yes.
P: Ok, can you tell us, what is a Hamiltonian path?
Me: (I was shocked at first, that they targeted the theoretical part of Data-Structures, I was a bit fumbled)
I don't know the exact definition, but I know it's like Spanning Tree.
P: No no, they both are entirely different.
After fumbling a bit,
P: If you don't know it, you can tell us, we can move on to other question.
Me: Yes
P: Suppose you're given a polynomial expression, what data-structure would you use for storing it?
Me: (Me moving over to whiteboard) I'd definitely not use an array. I'd go with vecor...as I'
P: (Interrupts me) I don't know vectors.
Me: (This got me nervous) Ok, I'd switch to using a Hashmap, as I
P: (Interrupts again) I don't know hashmap either
Me: (They were looking for specific answers only), I'd use a linked list.
P: Yes, that's what I know.
P: Now suppose there' another polynomial expression, and you have to add both, how would you do it?
Me: I'd keep Linked List in sorted order for this and explained further on the whiteboard. (they seemed ok with my solution)
Me: But this structure would face issues if I'd have to lots of issues in insertions, and if possible, I'd have used Hashmap (No reaction from their side)
P: I see that you've selected Machine Learning as your field of study.
Me: Yes, during filling my form, this was the most viable option that I understood, the other two fields I wasn't much sure about.
P: What do you know about clustering?
Me: (Shocked at first, as I wasn't ready for ML questions, but since I knew about clustering, I took a go at it), clustering, how it's an unsupervised form of learning
P: Ok, so you know. Suppose that you have a straight line and two cluster centers, C1 and C2, how would you minimize it. (These were the exact words, I didn't understood the question at all)
Me: Sir, I don't undertsand what you're trying to ask.
P: He repeated the same line again.
Me: (I drew a slanted x = y line on board)
P: No no, a straight line
Me: Like an X-Axis
P: Yes
Me: (Drew x-axis, and two cluster centers)
P: Yes, now minimise it.
Me: By minimise, do you mean optimizing it?
P: Yes, yes that's optimising (as if he was trying to say this word all along)
Me: Sir, I don't know much about this.
P: You'd take Difference between points
Me: Yes, I'd take dataset points and take difference between points and centers, and take a square again, to keep values positive, ahhh
P: But how do you know that you've minimised the cluster centers? How do you verify it?
Me: I don't know, Sir.
That's it. I don't have any chance in it. But learnt a thing or two about IIT Interviews, and shared it.
