Google L5 Interview Chances?

Intuit / Eng
SinghπŸ‡¨πŸ‡¦

Go to company page Intuit Eng

PRE
Adobe
SinghπŸ‡¨πŸ‡¦
Mar 5 37 Comments

R0 : Behavioural (Aced it)

R1 : delete even nodes from circular Linked List (Made mistakes but eventually solved it, many hints taken), swap nodes in pairs (did it with my eyes closed 😎)

R2 : Given a list of cards find the biggest straight in the cards (Interviewer was clumsy, her laptop lost charge and it took her 15-20 minutes to get started, I gave an unoptimized solution and then with a hint got to the optimal solution)

R3 : Design A Typeahead system (Luckily had studied it on Educative did my best here)

R4 : Build a MST, follow up built it without sorting the edges (Completely bombed πŸ˜«πŸ˜΅β€πŸ’«)

TC : 168k CAD
YOE : 7.5

#engineering #software #swe #google #faang #interview #sde #L5

94 PARTICIPANTS SELECT ONLY ONE ANSWER
VOTE VIEW RESULT

comments

Want to comment? LOG IN or SIGN UP
TOP 37 Comments
  • Amazon
    azrashk

    Go to company page Amazon

    azrashk
    Who the hell is building an MST in 30 mins at work ? This is such a ridiculous process.
    Mar 5 7
    • New
      denshSai2

      New

      denshSai2
      Yes the difference is now your neighbors is just all nodes not already in the tree, and you continue until n-1 nodes in the tree
      Mar 5
    • New / R&D
      hushhyena

      New R&D

      BIO
      Senior ML researcher
      hushhyena
      It's not that bad, if you work with graphs and trees a fair amount. If you don't, then it's a poor question, unless it's mainly just used to gain insight into how you operate when encountering a new problem.

      In the OP's case, you'd want to use Prim's algorithm, not Kruskal's algorithm, to avoid pre-sorting the nodes.

      Forget implementing this from memory, though, especially with optimal time complexity using Fibonacci heaps*. I'd honestly just tell the interviewer to grab the pseudocode out of Introduction to Algorithms, mainly to be cheeky. If someone is going to ask a question about a standard graph algorithm, then it should be permissible to talk about what to use and how it works without needing to implement it.

      * Knowing my luck, I'd get an interviewer who would say "BuT fIbOnAcCi hEaPs ArE sLoW iN pRaCtIcE" and ding me for suggesting them without realizing that they're honestly better than all of the fancier heaps that came after.
      Mar 6
  • This sounds like a no hire. Cant take hints as L5 and maybe if you had a redeeming coding round you would get extra interviews or downlevel. Sorry but I would say its a flat out reject
    Mar 5 5
  • What is an MST? Will you follow up and let us know the final decision?
    Mar 5 3
  • Google / Eng
    LeeJaeDong

    Go to company page Google Eng

    PRE
    Amazon
    LeeJaeDong
    Sounds like the typical loop: H in G&L, NH in all technical rounds.
    Mar 5 7
    • Intuit / Eng
      SinghπŸ‡¨πŸ‡¦

      Go to company page Intuit Eng

      PRE
      Adobe
      SinghπŸ‡¨πŸ‡¦
      OP
      What!! Why? It's pretty good I thought
      Mar 5
    • New / R&D
      hushhyena

      New R&D

      BIO
      Senior ML researcher
      hushhyena
      @SinghπŸ‡¨πŸ‡¦ From what I could see, the solution is pretty basic. If you want to impress enough for an L5/L6, then it helps to have a deep understanding of certain use cases. For example, you can talk about using low-memory approaches for estimating the frequencies of the most popular completion terms over a period of time and caching those, e.g., G. S. Manku and R. Motwani, "Approximate frequency counts over data streams", Proc. VLDB, pp. 1699-1711, 2012.

      I wouldn't expect a candidate to throw out something like that during an interview. If they did, though, then I'd be impressed enough to weight them as SH.

      Honestly, the best way to prepare for these types of questions is just to know your audience. Look at what Google is doing with their products. Think about what types of problems they have to solve to deliver those products. If you do that, then you'll come up with a majority of the systems design questions that they could ask. You can then prepare enough in advance.
      Mar 6
  • Looks like L5 just isn’t for meπŸ˜…
    Mar 5 3
    • New / R&D
      hushhyena

      New R&D

      BIO
      Senior ML researcher
      hushhyena
      @LC@Work Just prepare for it like you would a Ph.D. qualifying exam. Study your ass off for three months. Read the Introduction to Algorithms and memorize everything, including time and space complexities for various methods. Know the most efficient data structures for various types of problems. Then, knock those interviews out of the park.

      I hate leetcoding as a barrier for hiring. It's a terrible measure of someone's capability, as I've mentioned in other posts. I think it's a waste of time, especially if you've pushed production code that's used in a CMMI5 organization where lives are at risk if something goes wrong, have a well-established GitHub repository with our code, and so forth. But, right now it's part of the game. You might as well be good at playing the game and put yourself in the best position possible to negotiate.

      Eventually, someone will get the bright idea that leetcoding and brain teasers are terrible and move on to something else that will, hopefully, be better.
      Mar 6
    • Thanks @new. I just have my L5 loop in april end. So hopefully I do my best.
      Mar 6