Data science interview questions with answers • Hacker News Board

Re: Data science interview questions with answers

#71

Post by kvn_95 » Tue, Sep 15, 2020, 1:45 AM UTC

noodlenotes wrote: post #54

y42 wrote: post #32

Earlier quoted context omitted.

>> I prefer to ask candidates how they would approach solving a problem Word. Totally off topic, but: I work in the field of information technology since more than 20 years now, more or less. Not always the same focus, not always full time, but always IT related. I consider myself a good problem solver because of my self learning and analytical skills. I recently applied for a job as a BI developer. The interview con…

There's another possibility but only because you mentioned SQL specifically. People often use SQL as shorthand for "understand how to manipulate data," and if a new hire doesn't have this skill it can really set a data team back, so interviewers are touchy about SQL. It would be helpful if they clarified if it was the SQL syntax skills or the data manipulation skills they had a problem with. But generally I agree wit…

> But generally I agree with you that companies should prioritize problem solving skills over technical minutiae.

Unless, the company's goal is to hire someone with technical minutiae that they can pay the lowest wage possible :)

Re: Data science interview questions with answers

#72

Post by Jugurtha » Wed, Sep 16, 2020, 11:49 PM UTC

collyw wrote: post #68

ryndbfsrw wrote: post #64

Earlier quoted context omitted.

>> If you don't and just storm in talking about something something neural nets, they'll see it as a waste of time, won't bother explaining nuances >> Another major point is specifying the metrics. The real world metrics, not AUC or F1 scores. You need collaboration to get there, too. >> you need their help getting data, and their expertise to understand it. All this x100. And it's made worse as this humility isn't t…

Humility is discouraged in software engineering with job ads for ninjas and rock stars.

It should be taught. Lack of humility almost cost us a project. In that instance, humility unscrewed the project and unblocked 400k by simply sliding a sheet of paper and a pen to the client across a table, and asking them to draw the dataflow they thought we were going to build. A couple of boxes and arrows made it clear what the problem was. We drew the actual dataflow. The security person said "Oh, I thought... OK.. if it's like that then we're good to go". Legal said they're OK with that. Data people said OK. A dozen people were relieved.

The previous person on the project, although brilliant technically, thought they were "idiots who didn't understand crypto", as if it were the end goal. All it took was to keep quiet for a second and listen to what they had to say, and let them talk about what was problematic, instead of snark.

Re: Data science interview questions with answers

#73

Post by srean » Thu, Sep 17, 2020, 6:40 AM UTC

jmt_ wrote: post #67

bonoboTP wrote: post #65

Earlier quoted context omitted.

Statistics is not really in focus in most CS programs. Indeed I'm not sure where it is. Perhaps in applied math programs. Stats is kinda too dirty and realworld for pure math types, and in science programs and medicine it's usually just taught as a bunch of magic formulas to memorize and rules of thumb passed down from generation to generation without understanding. Perhaps physics departments do have both the necess…

This was my experience at a high ranking engineering public state school in the US. Stats is delegated to the applied math program usually (in fact, my degree was titled "Applied Mathematics & Statistics"). You can choose to concentrate in subjects like algorithms, operations research, fin stuff, statistics, etc. CS as well as other sciences had one required intro to prob & stats, but that's it outside electives. Fur…

> Stats is delegated to the applied math program usually

Which is not unsreasonable considering that these things arent really related to computers (although they happen to involve computation). I think its just an artifact of history and how things happened that ML is associated with CS and EE departments, but really its applied math, not a core CS topic like say compilers, formal languages and complexity.

Re: Data science interview questions with answers

#74

Post by collyw » Mon, Sep 21, 2020, 7:38 AM UTC

Jugurtha wrote: post #72

collyw wrote: post #68

Earlier quoted context omitted.

Humility is discouraged in software engineering with job ads for ninjas and rock stars.

It should be taught. Lack of humility almost cost us a project. In that instance, humility unscrewed the project and unblocked 400k by simply sliding a sheet of paper and a pen to the client across a table, and asking them to draw the dataflow they thought we were going to build. A couple of boxes and arrows made it clear what the problem was. We drew the actual dataflow. The security person said "Oh, I thought... OK…

I am actually starting to believe that (too) smart developers can actually be a hindrance in the wrong circumstances. We have a prima dona on my team and while smart, he doesn't make things easy to follow for the next dev.

Re: Data science interview questions with answers

#75

Post by Breza » Sun, Sep 27, 2020, 2:17 PM UTC

ryndbfsrw wrote: post #27

I’ve worked in the field for 7 years now so not that long but long enough to build some heuristics. The best data scientists are just people who try to understand the ins and outs of business processes and look at problems with healthy suspicion and curiosity. The ability to explain the nuances of manifolds in SVMs is not something that comes into it outside these contrived interviews. I prefer to ask candidates how…

I completely agree! I run a data science department at a corporation and I give applicants real world data and ask them to answer real world questions. "Here's a dataset of the dates we sent an email and the open rate. Does one day of the week have a higher open rate than the others?" People have to know how to turn a string into a date, turn a date into the day of week, and then apply statistical reasoning and then give me a one sentence answer. That kind of thing is far more common at work than explaining the kernel trick.