Interviewing with Microsoft...

[warz]

So, I've been in contact with a Microsoft recruiter since 2007, and have been waiting for my chance at an interview. Now that I've got Nortel on my resume, Microsoft has offered to have me come in for an interview. Typically I do very well in interviews, so I'm excited for the opportunity.

Newby, maybe your dad could give me a good word!

[Newby]

My dad's on his way out. I doubt that. That, and you're in line with MyndFyre for the same reason.

[Skywing]

For what position are you interviewing for?

[warz]

Skywing, I'm unaware of anything more specific than just Software Development Engineer. They've spoken to me about several of their projects that they've got going on, inlcuding something called Photosynth, as well as Silverlight. I doubt I'd be put on either of those projects since they're both well off the ground. I won't really know much more until this upcoming Thursday when I go into their office.

Newby, hehe I was just kidding. I've been plugging away at this recruiter for about a year. I know I'll do fine.

[while1]

Skywing, I'm unaware of anything more specific than just Software Development Engineer. They've spoken to me about several of their projects that they've got going on, inlcuding something called Photosynth, as well as Silverlight. I doubt I'd be put on either of those projects since they're both well off the ground. I won't really know much more until this upcoming Thursday when I go into their office.

Newby, hehe I was just kidding. I've been plugging away at this recruiter for about a year. I know I'll do fine.

so i herd u liek mudkip?

[warz]

i've got no clue what mudkip is :p

[Skywing]

warz: Which team, or do you not know?

[warz]

Well, like I said - I don't know, yet. I'm sure by this upcoming Thursday I'll know everything.

[Sidoh]

My friend is flying up to Washington to interview for an internship position this weekend.  I'm not sure if she'll be in contact, but I'll definitely be asking her how it went if she is.

I think Microsoft is one of the companies that sticks interviewees with moderately tough algorithms problems.  I'm guessing you've been reading up on these sorts of things, but here's a good example:

You have a car with an empty tank that you can move to any point on a circular track with no effort.  Gas cans which have an arbitrary amount of gas in each can (e.g., one can have 1 gallon, another can have 100 gallons) are placed around the track in arbitrary positions.  The total amount of gas in all the cans is just enough to get you all the way around the track.  Give an algorithm to find the point at which you can start your car such that you'll make it one lap.

The naive algorithm is O(n^2).  There's a fairly clever solution that allows you to get it in O(n).

This was a question that was asked by Microsoft a few years ago.  An alumni of the CS department here who now works at MS was asked that during his interview, apparently.  I wasn't sure I recognized his name, but if I remember correctly, he also interned there for a summer or two before he graduated.

[iago]

My friend is flying up to Washington to interview for an internship position this weekend.  I'm not sure if she'll be in contact, but I'll definitely be asking her how it went if she is.

I think Microsoft is one of the companies that sticks interviewees with moderately tough algorithms problems.  I'm guessing you've been reading up on these sorts of things, but here's a good example:

You have a car with an empty tank that you can move to any point on a circular track with no effort.  Gas cans which have an arbitrary amount of gas in each can (e.g., one can have 1 gallon, another can have 100 gallons) are placed around the track in arbitrary positions.  The total amount of gas in all the cans is just enough to get you all the way around the track.  Give an algorithm to find the point at which you can start your car such that you'll make it one lap.

The naive algorithm is O(n^2).  There's a fairly clever solution that allows you to get it in O(n).

This was a question that was asked by Microsoft a few years ago.  An alumni of the CS department here who now works at MS was asked that during his interview, apparently.  I wasn't sure I recognized his name, but if I remember correctly, he also interned there for a summer or two before he graduated.

Aha, the classic Interview 2.0!

[Sidoh]

The kinds of questions like the one I gave are a testament to how strong your problem solving/algorithmic skills are, which is vitally important in a number of roles at companies like MS.  The ones in the article you linked seem like useless "think outside the box" questions.  I've heard that some interviewers ask these kinds of questions, but at an info session with Google about a month ago, they said the interviewers were no longer allowed to ask those kinds of questions.

I don't at all think that question is useless.

[iago]

Well, whether or not that question is useful strongly depends on the job. I definitely couldn't solve it, and I'm awful at coming up with "good" algorithms. But, how often does that come up in day to do development? Generally, if you use anything but a simple algorithm, your code ends up more confusing without, in most cases, gaining anything.

Obviously that isn't true in every case, just like you said, but eh?

I like useless "think outside the box" questions better, since I generally already know the answer to all the common ones.

[Sidoh]

Well, whether or not that question is useful strongly depends on the job. I definitely couldn't solve it, and I'm awful at coming up with "good" algorithms. But, how often does that come up in day to do development? Generally, if you use anything but a simple algorithm, your code ends up more confusing without, in most cases, gaining anything.

I think I pretty much said the same thing in the first sentence.  If you're just going to be a code monkey, then strong algorithmic skills probably aren't that important.  If you're working on something like compilers, operating systems, data mining, parallel programming, pattern analysis, anything mathy, etc, your algorithmic skills and familiarity with data structures is going to make or break the interview.

I call BS on the "simple algorithm" statement.  Naive algorithms are often disgustingly worse than cleverer solutions.  This is true even in obscure problems that probably haven't manifested themselves enough to be written about.  If you're working with small samples, then it clearly doesn't matter.  However, when you're working with huge datasets (and there are plenty of jobs where this is the case), it's obvious that implementing naive algorithm after naive algorithm is going to produce a really crappy, slow product.

This is an aside, but: I enjoy finding algorithms for tough problems much more than implementing solutions.

[warz]

I've been through an interview just like this for Nortel. They ask questions that you don't quite expect to hear. I did very well at the Nortel one, though, because I was able to talk my way through the problem. That's all they want to figure out during these interviews. They want to see if a) you'll fit in with the team, b) you're able to do the job. The answer to these interview questions are going to make or break you during the interview if you're able to think out loud, and if possible draw what you're thinking on a white board, or something.

I prefer these "interview 2.0"s over other interviews because I think they're more my style. I'm much better at mapping my way to a conclusion than I am at coming up with algorithms on the spot.

Being able to code an algorithm, or whatever you're both stressing, isn't necesarilly of high importance during an interview if you can convey the theory behind the algorithm, or just explain what it'd do. Psuedocode is typically fine during these, and from the people I've spoken to that recently interviewed have told me that that's the case with Microsoft, also. It's Microsoft - they understand there's infinite amounts of references online, they're not going to expect you to be a human computer, but they'd like to know that you can use one to find the answers you're looking for.

edit: The thing that can be stressful about MS interviews is that they're not scheduled. You go to one person, and talk for awhile, and then get a 10 minute break, during this break they discuss if they think you have any weaknesses, or whatever, if the next interviewer will try to find out if this is true or not. Your interview can end early if they don't think you're a fit, or it could go on for 4-5 different interviewers if you're doing well.

[Sidoh]

I've been through an interview just like this for Nortel. They ask questions that you don't quite expect to hear. I did very well at the Nortel one, though, because I was able to talk my way through the problem. That's all they want to figure out during these interviews. They want to see if a) you'll fit in with the team, b) you're able to do the job. The answer to these interview questions are going to make or break you during the interview if you're able to think out loud, and if possible draw what you're thinking on a white board, or something.

I prefer these "interview 2.0"s over other interviews because I think they're more my style. I'm much better at mapping my way to a conclusion than I am at coming up with algorithms on the spot.

Being able to code an algorithm, or whatever you're both stressing, isn't necesarilly of high importance during an interview if you can convey the theory behind the algorithm, or just explain what it'd do. Psuedocode is typically fine during these, and from the people I've spoken to that recently interviewed have told me that that's the case with Microsoft, also. It's Microsoft - they understand there's infinite amounts of references online, they're not going to expect you to be a human computer, but they'd like to know that you can use one to find the answers you're looking for.

edit: The thing that can be stressful about MS interviews is that they're not scheduled. You go to one person, and talk for awhile, and then get a 10 minute break, during this break they discuss if they think you have any weaknesses, or whatever, if the next interviewer will try to find out if this is true or not. Your interview can end early if they don't think you're a fit, or it could go on for 4-5 different interviewers if you're doing well.

We were "stressing" over the importance of problem solving skills.  I don't think we were talking about the importance of how good you are at implementing an algorithm.

Sounds like a fun time.  Good luck on the interview.

I think I'll probably apply to Microsoft when they start accepting applications for summer interns (again).  The aforementioned friend who's interviewing this weekend is trying for an internship too, but apparently they accept applications in the fall and in the spring.  I already have three applications into other places (NASA, DoD, Lockheed Martin), and I think I'll probably apply to Google when they open up applications.

What's your reason for being interested in Microsoft, warz?  I mean I've definitely heard great things, but is there something specific that really attracts you to a job there?  I'm still kind of deciding where I want to work this summer  (and I'm by no means assuming I'll be accepted to all the places I apply, lol).