HondaCivet posted:

So how bad is it to give up on a coding challenge? I got one commit in before I got super stuck and then turned into a ball of nerves and couldn't think straight anymore. Thankfully it was over email so no one had to see me flip out, haha. I just told him that the problem seemed to be a bit over my current abilities. I know you should normally just do as much as you can but I just couldn't because I'm terrible.

astr0man posted:

And yes there are things like Fizzbuzz where the interviewer might actually be expecting you to answer it correctly, but in general it is more about just seeing how you think.

Coca Koala posted:

About six pages ago I posted saying that I had gotten a job offer for 85k in the Mountain View area and asked how that stacked up against other salaries for the area, and whether or not I should ask for more money. The general advice was that it's not a terrible offer, but that I should ask for more money because asking for more money has literally no downside.

I went and read the Kalzumeus blog post on salary negotiation, and I responded to the offer today asking for another 10,000 in salary, expecting to do some negotiating. They accepted my counter offer immediately and sent me a revised offer letter within three hours.

If you get an offer, ask for more money. There's no downside to asking for more money. And read the stuff linked in the OP, because it's for real helpful.

Strong Sauce posted:

Probably could have asked for more if they were quick to agree, but congrats!

Generally the thing is to ask for more money and if they say, "no" you can decide if you want to work there or not. It's not as though they'll rescind an offer to a qualified applicant.

Which is generally why you shouldn't set a price on yourself because if you had originally given $85k as your expected salary, you look like a jerk asking for more.

HondaCivet posted:

It was a bit more complex than FizzBuzz. It was taking a CVS file and dealing with the input which is stuff I don't exactly do every day.

Surely I'm not the only dev who's had a problem with nerves when it comes to interviews? Does it just take more interview experience to get over it? Is there a way to get help with this stuff besides flunking dozens of interviews for the sake of experience?

Brannock posted:

How much experience do you have Coco Koala?

HondaCivet posted:

It was a bit more complex than FizzBuzz. It was taking a CVS file and dealing with the input which is stuff I don't exactly do every day.

Surely I'm not the only dev who's had a problem with nerves when it comes to interviews? Does it just take more interview experience to get over it? Is there a way to get help with this stuff besides flunking dozens of interviews for the sake of experience?

HondaCivet posted:

I got hung up on something that turned out to be really dumb. I did find out how to read a CVS file but I got confused about what the method was returning. Then I got too panicked to actually figure out the issue when my stuff kept breaking.

HondaCivet posted:

It was a bit more complex than FizzBuzz. It was taking a CVS file and dealing with the input which is stuff I don't exactly do every day.

Surely I'm not the only dev who's had a problem with nerves when it comes to interviews? Does it just take more interview experience to get over it? Is there a way to get help with this stuff besides flunking dozens of interviews for the sake of experience?

HondaCivet posted:

I just feel bad because it seemed like a really cool job and I probably could have gotten through it had I had my wits about me. Oh well, at least I have some experience now.

JawnV6 posted:

I used to write emails to my mom when I'd get stuck, explaining the problem. Having to explain everything to someone nontechnical would often root out whatever faulty assumption I was making.

It's called "Rubber Duck Debugging" elsewhere, but when you're hung up on something like return type mismatched between your head and what's actually happening it's a useful tool. I also credit it with finding when I'd called the method OnPiant and was tearing hair wondering why it was never called.

ManoliIsFat posted:

smokyprogg posted:

So, apparently non-CS GPA isn't a dealbreaker, as I have an interview with Microsoft tomorrow.

That being said, oh god i've never done an interview for anything cs related gently caress.

I feel firmly confident with implementation and using of the structures/algorithms in the 2nd post, and I am fine with coding. This is for an undergrad intern position. Anything in particular I should prep for past the aforementioned?

bonds0097 posted:

Are you interviewing for PM, SD or SDET? And is this first-round or final round in Redmond?

smokyprogg posted:

First, and not exactly specified yet for which position. I indicated I would prefer SDET > SD > PM on the application.

astr0man posted:

Don't forget he's going for an undergrad intern position. They might not actually expect all that much technical knowledge.

Pollyanna posted:

How would the whole GPA thing work if you don't actually have a CS-based degree? Would an Engineering degree suffice?

wolffenstein posted:

I'm looking for a critique of my resume and social profile. I haven't been contacted by a recruiter in about a month, and even when recruiters were contacting me, I never got a coding challenge or an interview with a manager. I'm looking for a junior iOS and/or Ruby on Rails position.

Website, Resume, GitHub

Good Will Hrunting posted:

I know it's been said here a lot that the whole "X+ years" can be HR drone bullshit, but as someone with under a year of experience how do I approach this?

them posted:

In Java, C or C++, show me the largest 100 numbers in an unbounded stream of numbers.



I encourage you to write any limitations or assumptions you make about the problem or data set, etc.

me posted:

First assumption: We're using floats.

Second assumption: there exists a function called "insertionSort(float *nums, float toInsert, int listSize)" that performs an in-place descending order insertion sort, since insertion sort is a trivial implementation I won't waste time writing out. All it does is iterate over the existing list until it finds an element less than the input, shifts everything to the right, and places the input in that slot. If it's smaller than every element in the list it appends it to the end. The variable listSize should be at least one less than the actual array size since to insert into the list it needs to grow.

Third assumption: there exists a function called "float pollStream()" that returns the next number in the stream (if the stream is time-sensitive, we can assume it's buffered in another thread or something).

code:

int main( void ) {
    float numList[101];
    int arraySize = 0;
    while(true) { // Unbounded, so it goes forever
       float toInsert = pollStream();
       insertionSort(numList, toInsert, currSize);
       if(arraySize < 100) {
           ++arraySize;
       }
    }
}

The reasoning for this method is that insertion sort is O(SizeOfList) for an almost-sorted list. Since we're polling numbers off the stream all the time, the list should *always* be almost sorted (except for the newest element). Why is the array of size 101? The last slot is a "buffer" slot. If the element happens to be less than every other in the list, it gets inserted into the 101st spot, which we can happily disregard as junk data. Same if we find a new number in the max range, everything gets shifted to the right and the previous smallest number is now junk. Obviously the first 100 numbers are always the largest 100, hence the arraySize increment until it equals 100.

We could actually probably get a performance increase by using a linked list and simple inserting the new element to the middle rather than shifting all the data over in an array, but I figured a linked-list implementation would be non-trivial for a 30-minute question.

Of course, this just maintains the data. In a real scenario you'd want to print arraySize number of elements, or call the insertion function when you need it and then do something interesting with the data, rather than using a silly unbounded while(true) loop.

Jsor posted:

I guess part of my problem is that I think at an algorithmic level more than a code level?

quote:

I did mention the linked list in the post-script :p.

quote:

As for the heap, I guess. I thought of it, since O(n) goes to O(log n) but a heap really seemed like overkill (unless that was a joke). I didn't want them to think I was overengineering it.