#!process zon reject

1 messages · Page 1 of 1 (latest)

rugged kelp
#

how did you do in interview and when

vagrant pewter
#

interview was yesterday
i think i did fine
LP’s was decent. Coding was design based. I told him my approach, we had a long discussion. I was asked to then write a function and then we discussed more. Then we ran out of time!
i might have messed up somewhere!😞

static fjord
#

damn
sorry man!!

rugged kelp
#

do you require sponsorship by chance

glossy lotus
#

were you international

digital rock
#

There are many internationals who got amazon, dont think that matters

rugged kelp
#

i just thought that many internationals who did good didnt get the offer so it might matter a slightly

hasty timber
#

Damn is getting rejection that soon normal?

rugged kelp
#

im worried about me tomorrow this is crazy

solar gull
rugged kelp
#

timeline?

vagrant pewter
#

I am an international student. I don’t think that should matter considering 3 of my international friends got offers as well

graceful spruce
#

hey sorry to hear about the reject :/ what kind of system design question did you get if you dont mind sharing?

#

also when did you complete the OA?

vagrant pewter
#

OA-> 27th Dec
Interview Email->27th Jan
Interview-> 4th Feb
Reject-> 5th Feb

#

Question was something like:
predict the next word given last 3 words in a sentence
Predictions are already stored somewhere
Decide what kind of DS you want to store in
And given a sentence what will be the prediction
Word: I
Prediction; am
Word: I love
Prediction: to
Word: I like to
Prediction: eat
Word: I love to
Prediction: fish

Prediction will be single word
Predict based on last 3 words in a sentence

velvet ginkgo
#

How did you approach this question?

graceful spruce
#

seems like a binary search question

#

Trie question?

velvet ginkgo
rugged kelp
#

It’s only based on three words probably not trie?

graceful spruce
vagrant pewter
#

I gave a trie approach

vagrant pewter
graceful spruce
#

was your interviewer helpful?

vagrant pewter
#

He didn’t give any hints explicitly
Just kept asking questions
But he was really sweet

vagrant pewter
rugged kelp
#

Idk I think trie might be the approach but since it’s based on the last three word not everything right

#

This is a pretty hard question to be honest if trie is not the solution

#

You could have I like to fish and I like to hunt and how would you go about that

vagrant pewter
#

Yes correct
What i wasted my some time was I assumed i had to predict multiple words
But output was supposed to be single word

rugged kelp
#

Did the interviewer agree w you when u said trie

boreal veldt
#

Wait I don’t think I understood the question that was asked to you. Did you have multiple prediction “words” for the same 3 input words? I.e. I like to - eat and I like to - drink?

#

Or was it a 1:1 relation?

rugged kelp
#

Like is it base on frequency of one word appearing after three words?

#

Otherwise you would have no way to decide between two words with the same last three words

boreal veldt
#

And if it was a 1:1 as in 3 words could lead to one unique word, you shoulda used map instead of a trie

#

I don’t think it was as simple as a hash map tho, I think op would have been able to come up with that sol no matter how nervous

rugged kelp
#

But if you use map for a sentence with 9 words you would have at 7 entries

#

I think trie with frequency stored might be the go

boreal veldt
#

I still don’t follow the question lol

solar gull
#

Just ask gpt

vagrant pewter
#

For any given 3 words there was only one possible prediction

#

I initially did say hashmap but I wasn’t able to exactly explain. Once I explained Trie approach interviewer asked me a couple questions then asked me to start coding

solar gull
#

Most optimal was hashmap

#

For small inputs

#

Idek

#

Tbh

spice vapor
solar gull
#

Like if ur wondering so much just ask gpt

spice vapor
#

since there is only one possible prediction for any 3 words I think map is sufficient

#

if it was many possible predictions then trie

solar gull
#

Scalable solution is trie but yea

#

Since its small

#

Hashmap better

#

But tbh idk if that wouldve mattered in interview right?

#

Theyre very similar anyways

#

Im efficiency

spice vapor
#

let’s just end it at interviewer rng dogesmile

vagrant pewter
#

Yes I did mention Hashmap then I got confused and maybe thats where i fumbled and lost it

steep blade
#

I’m so confused

#

Where is the solution pulling “eat” from

#

What is the question even asking 💀💀

solar gull
steep blade
#

or like a huge dictionary of sentences

vagrant pewter
steep blade
solar gull
#

u just match

#

the input

#

to an output

#

literally

steep blade
#

Yeah … this question is kinda ass, but I think trie is the most intuitive negl

solar gull
#

trie is intuitive and the "smartest" approach that shows DSA knowledge

#

but a hashmap wouldve prob been better

#

regardless

#

i think this was interviewer rng

steep blade
#

The issue though is like, if the only input is “I love” how can you predict “to”

#

Since it’s not stored in your trie yet

solar gull
#

dictionary keyone

#

"i love"

#

dictionary value one

vagrant pewter
#

yeah i think so this

solar gull
#

"to"

#

literally thats it

#

tbh idk why this is difficult to understand maybe im oversimplifying it or not getting it

vagrant pewter
#

I just initially missed that there was only one output
and messed up this basic solution

solar gull
vagrant pewter
solar gull
#

maybe thats what the red flag was

vagrant pewter
#

Yeah correct
Hence the rejection in a day i suppose

solar gull
#

cuz ngl that is a trivial very trivial answer if thatsa actuallyt the answer

#

yeaaaa

#

oh well dude

#

im sure it was just nerves

#

cant think right when ur nervous

steep blade
solar gull
#

key value pairs

#

the key is the sentence up until the word

#

then the value can be the predicted word

#

literally thats it

steep blade
#

Yeah but if your map doesn’t have the word

solar gull
#

theres no real rpediction its just mapping a value to another value

solar gull
#

its set

steep blade
#

How do we know tho

solar gull
#

it says in teh question

#

that the predictions you already knoww hat theyre gonna be

#

just based on alst 3 of a sentence

#

so just exttract teh last 3 in a sentence

#

and compare them againsdt u hashmap of ketys

#

trivial

#

id cry of joy if i got this as my question

#

i havent even prepped lc and i could solve it

steep blade
#

So they give us the prediction … and we’re trying to find the prediction

#

Oh I’m regarded

#

I didn’t even see the “predictions are stored somewhere”

solar gull
#

yes vro

#

leetcode easy

steep blade
#

Lmao no way

#

I didn’t think that would be the case, or else this question was dummy easy

#

I thought we were building our data structure along the way

solar gull
#

doesnt seem like it

steep blade
#

So we would have to mark the start node of the word in the tree

solar gull
#

perhaps the reject

#

was due to overcomplicating it

#

with tries

steep blade
#

Yeah def

#

Idet trie is wrong

solar gull
#

but idek i mean its nto like it was a bad answer

steep blade
#

Like I would answe r it both ways, describing the + and - of each way

solar gull
#

its nearly the same efficiency

#

yea

#

maybe just red flag not being able to intuit hashmap approach

#

just nerves prob

steep blade
#

Yeah def

#

Unlucky

fallen furnace