?

Log in

No account? Create an account

Previous Entry | Next Entry

another job interview

I had an interview today at Y! with the Strategic Data Systems group, which (roughly) is the operations group that supports the people who crunch the web server logs. Tu Tran, who some of you may remember from AV ops, was one of my interviewers.

Some high (low) points of the interview:

"How would you determine who was the last person to modify a file using /var/adm/messages that is mode 777 (read/write/executable (rwx) by all)?"

Me thinking "Huh? I have no idea." After a while, finally saying "if there's a command to do this, I don't remember." The answer, which surprised me, was to change its mode to 700 (rwx by owner) and when someone tried to modify it, a message would appear in /var/adm/messages indicating there had been an attempt to modify the file by a given user. Honestly, I had never known that, and as many times as I've looked at /var/adm/messages, I don't recall seeing it.

"What's the difference between == and eq in Perl?"

Me thinking "I've used this, but I don't remember." The interviewer gave a clue about how one is used for one type of data, and the other is used for another. I spent a bit more time thinking about it and vaguely remembered it had something to do with strings and integers, so I said that one is used for strings, and the other, integers, but I didn't remember which. It turns out that if you use == on strings you get a warning (if you run the program with the -w option).

"How would you implement file completion in a shell?"

Me thinking "Huh, I have no idea! I've never done it; it has never come up. This question is really open-ended." After contemplating whether or not I could BS my way through an answer, I decided it probably wasn't a good idea, so I said that I had never done that, and that I'd probably have to think it over and study some examples.

I did give some good answers, but apparently not enough, or the not so good ones outweighed the good. With about three more people to talk to, the guy who phone screened me came in to tell me that based on the previous interviews, they decided to pass on me.

I did, however, learn a couple of useful things that afternoon.

I asked my phone screener what the interviewers are really looking for in a candidate. He said that a lot of the time, they're looking for people who know the answers to "geek trivia" (which I generally think is the category most of these questions I seem to get fall into). They're also looking for people who seem really excited about the company. The way he made it sound, this type of excited could be described as hyper — he gave an example of how he'd had several cups of espresso and was really wired. (Reminds me of what andybeals wrote sometime back when I mused on using drugs to handle interviews better.) He also said that there were some times that he worked incredibly long hours trying to fix problems (not caused by him), which is something I definitely want to avoid.

Seeing as I didn't think the job was a good match to begin with, I'm not too disappointed. Perhaps a bit disappointed that I didn't remember some things that I actually know.

Comments

( 14 comments — Leave a comment )
slfisher
Apr. 15th, 2006 11:27 am (UTC)
See, I think that's nonsense. I wouldn't want someone who knew trivia. What I would want is the person who would say, "This is where I know to look it up." Because, you know, things change, and I want someone who can learn and grow.
bryant
Apr. 15th, 2006 11:35 am (UTC)
Speaking as a guy who hires tech people... when I hear the classic "well, I don't necessarily know the answer to a given question immediately but I can look anything up," it raises a slight flag. I do need that quality. But I also need someone who's had enough experience with whatever we're doing that they can see problems coming and troubleshoot quickly.

And there's a difference between someone who's been tuning Apache for five years and someone who knows how to find information on Apache tuning, no matter how quickly the latter can pick up the knowledge. I worked at Sun for three years and by the time I left I could tell why a Solaris boot was failing by the rhythm of the lines scrolling up the console. ("Oh, there's the delay -- it's failing the NFS mount.")

So if someone has deep experience in a specific area /and/ they tell me they learn quick, that's awesome. I just get a lot of people who have no deep experience but a good line about researching problems, and that makes me wary.

Oh, and that first interview question is bogus. Their answer will tell you who the next person is, but not who the last person was.
agrimony
Apr. 15th, 2006 02:28 pm (UTC)
I was just going to make that same point as Bryant on the first question. It's a semantics thing, but yeah. That's waiting to see who the next person that attempts it is, not the last person who touched it. If the ls -l info doesn't say who the last to modify it was (or if it's a generic account), then I suspect that you're hosed unless you've got some kind of logging of command execution going on.
gregbo
Apr. 16th, 2006 06:31 am (UTC)
Speaking as a guy who hires tech people... when I hear the classic "well, I don't necessarily know the answer to a given question immediately but I can look anything up," it raises a slight flag. I do need that quality. But I also need someone who's had enough experience with whatever we're doing that they can see problems coming and troubleshoot quickly.

For those unaware, I worked with/for Bryant for a while at AV.

A bit of background: When I was first contacted by the phone screener (he was the one I mentioned earlier who brought an O'Reilly reference manual to an interview), our conversation touched on a number of issues. He asked me what I had worked on at AV/Overture.(Apparently, he's recently inherited something having to do with Overture). After telling him about it, he asked a few more questions, then expressed some surprise that I didn't have a job right now, despite having a master's, etc. I told him about some of the things I've written about here (forgetting things during interviews), and also how I felt that age was a negative factor in terms of me getting jobs. The long and short of it was, he still thought that I was a candidate. So I told him to contact (or have someone he reports to contact) David Bills (my grandboss at AV once business ops eng was put under him), who would tell him all about me, what I'd done, what I thought about what I'd done, and whether he thought I'd be a good match. I said I would abide by whatever he said. I also told him that being out of work, I wasn't exactly in a position to turn down honest work.

I hadn't yet actually seen a full job description. When I got the invitation to interview, the email included the full job description. I hadn't actually applied for this job; it was something the Y! recruiters passed on to the group that interviewed me. In fact, I remember seeing this position and passing on it because (1) it involved some things I didn't know about and (2) it wasn't really what I was looking for. So I figured I'd go in anyway and just see what they had to say. I thought if anything, it would provide some extra interview practice.

My phone screener was the last person to see me that afternoon. In addition to discussing the "geek trivia" aspect of the interview, I asked him if he (or anyone) had ever contacted Mr. Bills. He said he hadn't contacted him because the thought of talking to someone in an executive position intimidated him. Possibly, someone above him in the management chain talked to Mr. Bills, but he didn't know.

As an aside, some things he said sort of puzzled me. He said he hadn't actually met Mr. Bills or Manny (the planning and logistics head), but he vaguely recalled hearing their names, and seemed to think they were based in Pasadena (Overture's HQ). AFAIK, both of them still work in Sunnyvale, although Mr. Bills is still on the road a lot.
bryant
Apr. 16th, 2006 02:29 pm (UTC)
Note that I'm more responding to slfisher's comment than the original post -- in my experience, you're someone who knows what he needs to know and knows how to find out what he doesn't know. I.e., the kind of guy I'd hire.

If your phone screener was nervous about talking to your reference, there's something wrong with his approach to the hiring process. That's what references are for.
gregbo
Apr. 16th, 2006 09:55 pm (UTC)
Thanks for the compliment. :)

If your phone screener was nervous about talking to your reference, there's something wrong with his approach to the hiring process. That's what references are for.

Actually this isn't the first time something like that's happened at a Y! interview. When I interviewed with the abuse group, the project manager was my initial contact/phone screener. During the course of our conversation, I mentioned David Bills and suggested that he speak with him to get more background information about me and the projects I'd worked on (especially since the project manager was fairly new and not very familiar with how AV and Overture became part of Y!). He said that he'd rather interview me first, which struck me as kind of odd, for the same reasons you state.

In the past, using references and recommendations seemed to work (not just for me, but for everyone everywhere I'd ever worked). It seems like they're much less trusted now. Maybe too many employers have been burned too badly. Also, as noted in a blog post on Google's hiring strategy, they don't use hiring managers.

One of the reasons I follow the click fraud issue so closely is because I want to understand how as smart as Google is, they didn't do enough to at least limit click fraud when AdWords and AdSense were first released. I don't understand how they couldn't have realized that they would have a serious problem with click fraud. I realized it, and I'm not as smart as many (maybe even most) of them, but I didn't have resources they had to do anything about it. The strategy blog post also notes that they ran simulations to justify their hiring processes. Perhaps they also ran simulations to assess their risk for click fraud. My general opinion is that they didn't quite take into account the economics of the Internet, where accounts, traffic, etc. can be easily and cheaply created, identity can be easily concealed, and so forth.
gregbo
Apr. 16th, 2006 08:07 am (UTC)
One more thing ...

I was thinking back to around the early summer of 1999, when I'd been dancing for about a year and a half. I was testing out of Bronze 2 in the old Starlite syllabus. There was a Nightclub Two Step pattern called Change of Places to Underarm Turn Left Loop I had to do with my teacher. It's something I used to do at parties regularly. So I started to do it, but couldn't remember how. It was really strange. I tried to think of what to do, but just couldn't, so I asked the tester to just skip it and move on to the next step.

A couple of days later, the tester asked me and my teacher if I would do the step again. He said he'd seen me do it at parties many times, and thought I could do it; he just wanted to see me do it. So I did, and got credit for it on the test, and went on to pass everything else including the freestyles at the medal ball.

To relate this to the present situation, I was sort of hoping that someone like David Bills could speak up for me, in the same way the tester did. Someone who knew what I was capable of, even though I may not have shown it at an interview or a test. It seems that in this age of heightened competition in the computer industry, people trust the test they give more than what someone else who's worked with them in the past might say about them.

railmeat
Apr. 15th, 2006 03:29 pm (UTC)
I agree the first question was bogus. I think you would only see a message if the permission change were done via sudo, right? Also I don't see a /var/adm on the FreeBSD machines I have access to. Maybe that was /var/log/messages?

The second question was more reasonable. It sounds to me like your answer was reasonable though.

I have never quite figured out the "excited about the company" thing. It is an explicit part of the interview questions HR sends out. On the rare occasion that I don't really expect people to be excited about Yahoo. We are a big company now. I would expect someone to know something about the job they are interviewing for, or at least show some interest. Maybe I don't think it is important because I don't get excited about much. Especially not Yahoo! now that I have been working there for a couple of years :-)

gregbo
Apr. 16th, 2006 02:28 am (UTC)
From the sudo man page:


If sudo is run by root and the SUDO_USER environment variable is set, sudo will use this value to determine who the actual user is. This can be used by a user to log commands through sudo even when a root shell has been invoked. It also allows the -e flag to remain useful even when being run via a sudo-run script or program. Note however, that the sudoers lookup is still done for root, not the user specified by SUDO_USER.

sudo can log both successful and unsuccessful attempts (as well as errors) to syslog(3), a log file, or both. By default sudo will log via syslog(3) but this is changeable at configure time or via the sudoers file.


I've never used this feature because I've never needed it.


ascendneworder
Apr. 16th, 2006 12:57 am (UTC)
Having an interview with Yahoo is great! It's a tough company to land one at. I'm sure they don't expect an exact answer for each question, but to see how you answer and that you answer with your best attempt with some knowledge in the area. You can write a thank you email back saying that you got nervous and actually knew some of the questions. Be positive, maybe they like you more than other candidates and you fit into the group better even if you miss a question.
gregbo
Apr. 16th, 2006 01:08 am (UTC)
You can write a thank you email back saying that you got nervous and actually knew some of the questions. Be positive, maybe they like you more than other candidates and you fit into the group better even if you miss a question.

They've already passed on me. I thanked the guy who gave me the initial phone screen. (No one else gave me their contact info.) I'll write more about the getting nervous part later.
cellio
Apr. 16th, 2006 01:29 am (UTC)
I don't care for "geek trivia" questions either. I want to know if the person can {design, code, debug, work out architecture... whatever's appropriate to the specific question). I'll certainly ask "how would you...?" and "what's the difference between...?" questions, but they'll be real, not trivia. The trivia will fall out by listening to the answers to the real questions, after all.

Context matters. If someone says he's been hacking Perl for five years, the eq/== question is fair. If he says he's done some Perl hacking, I'd pass on that. For Java programmers, I like to ask about the differences between interfaces and abstract classes (and when to use which); there's not a single right answer, but it can lead into a discussion of design in general and that's useful. (Similarly, I ask about extension versus delegation -- "is a" versus "has a".) But I don't really care if they can quote arcane details about Swing (though maybe I would if the job was in that specific area -- you see what I mean about context?).

(The first question they asked you is bogus, as others noted.)
gregbo
Apr. 16th, 2006 04:00 am (UTC)
I actually thought the eq/== question was fair. I just didn't remember how to answer it. I've used eq ... I just didn't remember how I used it.

When the question was asked, I thought immediately of two (unrelated) things: eq/equal in Lisp/Scheme and is in Python. (I was actually going to make a separate post someday about both of these someday and how they relate to 6.001.) I then tried to think of code I'd written that used eq. I couldn't quite picture anything. After that I was given the hint and eventually the answer.

Context matters. If someone says he's been hacking Perl for five years, the eq/== question is fair. If he says he's done some Perl hacking, I'd pass on that.

Fair enough, I suppose. Part of the problem is that even though I've worked on Perl for several years, it hasn't been continuous. For example, I didn't do any Perl programming at all from 1995-1997, but had done a lot from 1992-1995. Also, during the seven years at AV, I did the bulk of Perl programming from 2001-2004, but the job was not exclusively Perl programming. I had to do a lot of other things, many of which fell outside of a technical realm. This was something I was very concerned about, especially since I compared myself to other people who held the same, or similar titles as I did, but their jobs enabled them to focus more on engineering in an engineering context. (For example, I had to spend a lot of time explaining things like why cookies are not the same as people to people in sales and marketing who did not have technical backgrounds. So I had to spend a lot of time simplifying the explanation and making analogies.)

Also, after being laid off from AV, I haven't worked on a major Perl project. The people I worked with at Nominum dislike Perl a lot and prefer to use Python. I had to teach myself Python since I'd never programmed in it before. So I'm rusty in Perl. (OTOH, I don't really know much Python and have forgotten some of what I learned.)

I'm sort of caught in a difficult place because if I spend time on one thing, I'm not spending time on the others. I can't predict ahead of time what I'll need to know on an interview or in a job. (In fact, I didn't know I would need to program in Python at Nominum until the FMC project was canned and I was moved into the test and release engineering group.)
aelfsciene
Apr. 16th, 2006 06:00 am (UTC)
Wow, reading this whole interview process really makes me glad I'm not in the tech business anymore.

Sorry to hear it didn't go well, but if it's not a good fit, it's probably for the best. Good luck on future possibilities, though!
( 14 comments — Leave a comment )