Google Statistical Coding & Algorithms Mock Interview

Today I am interviewing Dan, a machine learning and data engineer, on a common statistical coding and machine learning interview question around calculating medians from a stream of numbers.
Here's the question we're tackling: www.interviewquery.com/questi...
Want to be featured in the next mock interview video? Apply here: airtable.com/shrdQrwKK7xxGLm6l
👉 Subscribe to my data science channel: bit.ly/2xYkyUM
🔥 Get 10% off machine learning interview prep: www.interviewquery.com/pricin...
❓ Check out our ML interview course: www.interviewquery.com/course...
🔑 Get professional coaching here: www.interviewquery.com/coachi...
🐦 Follow us on Twitter: / interview_query
Follow Jay on:
Twitter: / datasciencejay
LinkedIn: / jay-feng-ab66b049
Related Links:
Google machine learning engineer interview profile: www.interviewquery.com/interv...
Google data scientist interview: www.interviewquery.com/blog-t...
Google Data Scientist skills profile: www.interviewquery.com/interv...

[@juntaoxie7067]

isn't that insert for method 1 is O(n) instead of O(log n)? since slicing is O(n)

[@Sn-nw6zb]

Just use two heaps, left heap for lower numbers and right heap for higher numbers. And then you can make sure insert complexity would be O(log n)

[@consistentthoughts826]

I performed well in a case study discussion for a small startup after watching these sessions Thank you very much

[@omkarsunthankar2061]

Use two heaps min heap of all max values and max heap of all min values. Getting media. Is 0(1)

[@kshitizkhatri9733]

This was pretty good quality development and debugging. Loved the test driven development approach here.

[@qz1771]

I'm guessing the best answer would be a self-balancing binary search tree

[@basicmaths3443]

my interview is for 1 hr, if i take 34 min in solving a median question then ?

[@mdazimulhaque]

Great interview. Just a question though. Are statistical coding questions limited to Google or other companies also ask it in data science interviews?

[@iqjayfeng]

They’re pretty common across many data science interviews because of their two birds one stone approach.

[@mdazimulhaque]

@@iqjayfeng thank you for answering!

[@basicmaths3443]

Can I solve like this? Will the interviewer like my this solution?? lst= [ ] def median_cal(lst): print("The original list : " ,lst) lst.sort() n = len(lst) print("sorted list : ", lst) med_index = n//2 m = (lst[med_index-1]+lst[med_index])/2 if (n%2 != 0): print("Median : ", lst[med_index]) else: print("Median : ", m) return

[@LifeofTF]

Any BF deals for interview query?

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@spicytuna08]

I got this. would this satisfy the question? def calculateMedian1(list1): list1.sort() listSize = len(list1) medianIndex = int(listSize/2) median = list1[medianIndex] if listSize % 2 == 0: otherMedianIndex = medianIndex - 1 median = (median + list1[otherMedianIndex])/2 return median

[@basicmaths3443]

I solved like this. You need to pass a list as argument in the function. lst= [ ] def median_cal(lst): print("The original list : " ,lst) lst.sort() n = len(lst) print("sorted list : ", lst) med_index = n//2 m = (lst[med_index-1]+lst[med_index])/2 if (n%2 != 0): print("Median : ", lst[med_index]) else: print("Median : ", m) return

[@pigdogcheap1773]

I think he's confusing the Median with something else. A median should be a number in the given array . 2.5 and 5.t that he mentioned in both examples aren't the medians

[@TpBrass0]

Ugly code…

[@mingunpak1786]

I don't think he can actually pass the interview.

[@purnawirmanhuang5145]

agreed, he start coding before he explains, which is a big no no

[@Hassan.Wahba.97]

@@purnawirmanhuang5145 The algorithm was also all over the place. It could have been way much easier... he needs some software engineering classes to start with.

[@yawninglion]

I would fail him too. 1. He apparently didn't consider using heaps. 2. The way he handles bad input is dangerous. When the class is provided with bad input, simply returning none silently makes debugging hard.

[@harshmankodiya9397]

Thanks for confirming I don't have to watch the while vdo no😅😅