Tuesday, October 14, 2008

Yet Another Interview

An interview call makes you really excited - you think of all the possibilities and aftermath. After the process is over, you are drenched. It might leave you elated if you have done exceedingly well or tired/fatigued if you failed to make a mark. So why am I talking about all this?

After the Bloomberg offer, I send denial emails to few companies I was being interviewed with. Believe me, its hard to send in such emails because I understand how much effort has been put just to get that call. I was lucky I did not try for many and not too hard (Y! rocks in some sense none the less).

I finally applied to Microsoft and going to have an interview on 16th. Everything was going on nice and fine. Till this happened and perhaps I have not had good sleep till yesterday. I can go for a deep slumber now as I don't have much hope from today's encounter :-|

I got an interview with Google. [Please do not ask me about questions - I cannot discuss, nor the emails;]. While I did not expect much preparation, I punched in several lines of code as this was supposed to be a coding exercise. I was doing fairly well with one fear that what if the question does not click me.

What I have realized is that if I know the solution well, I can jot down the code without making a lot of mistakes. Believe me, my code would show up with fewer compile time errors and hmm, a few runtime ones (those which are really really evil and happen only on full-moon days).

But when I do not know of the solution, somehow I am good at explaining them but if asked to explain as well write the code simulateneously, I falter. I perhaps manage to pitch in the overall picture correctly by picking up the grassroot problem; but when it comes to finishing it up - I just go eerie with design blunders. I am not sure what screws up my frame of mind. Is it the constant pressure which mounts every minute of the interview or is it that I have exhausted myself of that creative juice by not taking ample rest before the interview.

A lot of things can be spoken after the interview - the fact is plain & simple. There is room for improvement. I might have polished my coding standards and understanding of c++ a lot better. Given a known algorithm, I can write up the solution quickly and efficiently (something which again I need to work on; I found that I tend to do a lot of pre-optimization; something which brings in those runtime errors - there is a simple solution to this :: Hold yourself till you execute the code). I need to write in more and more code; code which makes me think while I write. I am more of a paper-pencil guy who likes to solve the problem separately and code separately. But this has to change. I don't know how! But perhaps more practice and more experience.

Now that I have begun with one down, lets see how I fair at the MSFT interview. I am not preparing for the same because I prepared a lot for today's interview. I want to take it all in the light spirit. Perhaps over-preparation spoils the party.

Summary:
1. Do not optmize before you execute. Let your code be verbose. You need to make that cut.
2. Relax, relax, relax
3. ?If the solution does not strike to you and you have to start coding any how, What should you do? - Try to find out.

Friday, October 3, 2008

Not just the pursuit, I am happy today

Well not just that I am sober, I am typing this at almost the stroke of the midnight making a lot of spelling mistakes and correcting them. So you know how difficult it is for me to keep it grammatically right till now (pardon me if I really made any).

I finally could score a job offer and look forward to join NYC for the next year. It's a dream to work in NYC. It reminds of so many things I have missed out and I have high expectations from myself as well as the firm I am joining.

I have time for almost Feb 9th to enjoy and make out something finally of this Masters. I know this degree gave me a lot of things to ponder and speculate. But still, the ambitions are not yet fulfilled. It is long road and I need to envisage it. This is perhaps the first step, or perhaps the first success of the few bold steps I have taken in the recent past.

For the die hard techie fans:

From now on::::

From now on::::

Whenever you compile using g++ and use inline ............................... yes u use inline to make your program faster;

Use g++ -O2 or -O3 ...... then only it takes into effect. AND, use gprof to actually measure it, otherwise there is no point in making a function inline - really absolutely no point at all.

Learn how to profile your code today before it is too late, regardless it is jprobe or gprof. I am sorry for the intoxication :)

SiteMeter