Technical art coding test query

Coding
,
#1

Hi there guys!

Had a question about coding tests. Are they common in your experience?

I recently did one for ea, 3 questions were python and 2 were c#. I managed to get 4 out of 5 done in the 100 minutes provided, but just missed the timer for the last one, figured out the bug in the code I had written a few minutes after :face_with_hand_over_mouth: ( I read the question wrong so had to restart the entire code half way through) :joy:.

Wanted to know what you guys thoughts about this? Do you think missing the timer for one will rule me out for the job? Im hoping thats not the case but I have never done a timed coding test before :man_shrugging::man_shrugging:. I figured TA interviews are not as strict as engineering tests (from my experience you have to get every question correct with the most optimal solution) but I really have no idea
:sweat_smile::sweat_smile:

Thanks all :blush::blush::+1:

#2

In my personal European experience coding tests are quite rare.
But maybe that’s just me :thinking:
Also, I most certainly never had a timed test in my life.

#3

At RSE we do a tech art test with some coding/shading questions but it’s not timed.

I’m not sure I see the value in a timed test. It’s not a realistic situation that i’ve ever experienced that I suddenly only have 100 minutes to solve 5 problems, especially all by myself.

Even with the non-timed test though it’s more about seeing what solutions they came up with and their style as opposed to a hard pass/fail based on “did the code run”… That’s just me and some of the people I work with, obviously every company is different and every interviewer within that company probably has some standards they’ve just arbitrarily set.

I think it also depends on your skill level. I’d expect a senior to be able to handle most problems with a solution at least, maybe not the most performant but it seems like in most cases, you either get good code or quick code :slight_smile: and in a 100 minutes you get quick code.

Sometimes these test just measure how good they are at leetcoding not so much coding or problem solving.

#4

I have done them in the past, am ambivalent about the value.

What I did was give candidates a set of descriptions and unit tests which told them if they had gotten the answer, and asked them to turn it around within a set time (I forget how long… ~ 2hrs?).

The tests themselves were not super hardcore CS problems, they were intended to provide a range of areas where people could display their coding style and general knowledge, not to weed people out. By definition, any code that passed the tests was “good enough”, and the bar was deliberately not set so high that most candidates could not do it at all. Instead, I wanted to see things like “does this person know when to use a generator vs a list” or “do they write a lot of unnecessary classes” or “do they know how to X algebraically instead of brute force.”

The key thing was to provide clear guidance about the nature of the tests, what would be looked at, and what would not.

I found it useful, but mostly as a way of broadly sorting people into low, middle and high tiers as coders so they could be interviewed appropriately. Sometimes you do need people to clear a skill bar – but in general, any evaluation method intended just to bounce people is pretty lame.

#5

FYI thought id let you guys know that I got the gig :partying_face: :partying_face: :partying_face:

2 Likes