• First of all, I would like to say BIG THANKS to Leedcode community for all the insights and intuitions which I gained by going through many solutions, interview experiences and compensation posts.

• Secondly, I want to limit this article to share my take aways with my first attempt at Google, I got offer from Microsoft for L-62, rejected at Amazon, Selected but reject (hiring manager / hr discussions) to release offer with Salesforce and ServiceNow due to fitment and budget (ServiceNow) issues.

I don't think I really have some important information to share which was not been already discussed on many other posts regarding all other interview experiences except Google.

Update: As per may requests on the comment I shared my microsoft experience here

Interview Process

0. Interview call
I got interview call through referral after 1 month of applying for 2 generic sofware engineer roles (full-stack, backend)

Note: If you are despiratedly wanted to give interview with Google, it is no way tough to get it, all you need to do is keep connecting recruiters through LinkedIn and share your leetcode / other coding platforms profiles with no of solved problems count the more practice you do the more chances you will be given opportunity.

1. HR Screening Round (Phone)
This is an obvious round HR will just try to see how much prepared you, I have a asked basic questions (4- 6) like

i. What is the run time complexity of Merge sort?
ii. Which sorting algorithm is best?
iii. What is the value of 2 power 22 (I told like square of 2^11 => 2048 * 2048 => approx 4million+)

2. Phone Screen Round (Google Meet)

• Problem: Variation of Prefix search and a follow-up, Trie based solution one can solve it if they learn Tries. Leetcode similar question here

• Result: Positive

3. Virtual On-Site - 1 (Google Meet)
In this round interviewer inital spend 10 mins in talking about my current work and project before jumping into problem

• Problem: Given directed acyclic graph find the maximum length for each vertext that posible from each vertex in the graph, I could not find this in leetcode but exact same quesion is in GFG.
  Example: You have edges a -> b, b->c, a->c, output should be: [{a, 2} , {b, 1}, {c, 0}]
  from a you can have a path of max length - 2, from b you can have max length - 1, from c you can have max length - 0.

• Result: Positive

4. Virtual On-Site - 2 (Google Meet)

• Problem: Its a kind of pin ball game where you will be given 2D matrix board with obstacles ('\', '/' ) and start and end position, you need to return the end point by travelling ball from start point. When ball hits obstacle it changes it direction perpendicular to the ball current direction towords the face of obstacle.
  Ex: ball going to Left -> Hit the '/' it turns to Down side.

  				 - - -
  				|/|/|\|
  				 - - -
  				| | |\|X
  				 - - -
  				| |\| |O
  				 - - -
• output: position of X i.e. (1,3),
  explanation: ball starts from (2,3) left direction, move to (2,2) then (2,1) change direction to UP, move to (1,1) then (0,1) change direction to right move to (0,2) change direction to down move to (1,2) change direction right move to (1, 3) you are at exit point X.
• Result: Negitive

5. Virtual On-Site - 3 (Google Meet)

• *Problem: Validate an equation (return true if the expression is valid).
  Possible literals: +, -, /, , a-z, (, ), =. all operators are binary.
  Example: a + x = b + ( z - ( a + w ) ) output: true

• Result: Average

Overal result: Negitive, So did not scheduled Googlyness round.

I have all rounds happened on different days spaning a week, from phone screen to on-site rounds I took 40 days time.

Insights, Observations, Take aways

1. All the quesions in all the rounds started with 1 liners with out interviewer writing anything on the coding document that they shared during interivew, which tells that How much important it is for google to resolve ambiguity, set the scope and constraints before solving the question.

2. I have asked clarifying questions and input formats, types, output format, constraints for every question except my "Virtual On-Site - 2" where I got Grid pasted by interviewer and explained what he wanted to find. So I would suggest all of you who targets google it is important to make sure you frame the question as quickly as possible.

3. Google conduct very strict timeboxed coding interivews we typically gets 10-15 mins to write code after resolving ambiguities and brainstorming solution so it is very important to practice more mock interivewes, consise thinking and clear demonistration of thoughts.

4. In my case, I solved every question in all of my rounds, but in "Virtual On-Site - 2" I took 35 mins to completely understand and Identify the solution due to little confusion about Directions and its reflection interviewed found where I'm thinking wrong once he told me about my mistake I could solve it and implemented but in just 5 mins with logically correct but syntactically few mistakes. I choose Recursive approach later interviewed pointed out that I could have solved it in iteratively and I missed some return keywords in the Recursive approach I took, He also pointed out representation of direction could be enum, my confusion and mistake about understanding directions-part siting all these concerns "Virtual On-Site - 2" resulted negitive feedback.

5. Google has changed process little bit for upto L3 roles they are giving cadidate choice to choose either 3 coding rounds or 2 coding rounds and System design.

6. IMO, writing clean code and not struggling much on logic when you are into implementation is very important, they are keen in looking at access specifiers like public, private (if you use java, c#), namings etc. While I was implementing solution based on stack for my "Virtual On-Site - 3" I struggle little bit when writing if and else statements, push and pop and keep on going to back to input string debugging and proceed in writing logic that drags me into Average rating side.

7. Do lots of practice with mock interivews, I suggest avoid mock interviews with strangers find a partner whom you trust and identifies your weaknesses and provide you feedback to improve upon. This is some thing I missed in my case when you do with strangers you can not actually track or underastand your improvements.

8. Do not take too much pressure on the 45 mins I find the questions are relatively reasonable enough to solve it in 45 mins given a candidate have decent practice and consise thinking.

9. I have talked what ever comes in mind when I brain storm solution in the verge of thinking loud, but I realised I should have kept a filter and validated for a second before spilling out every thought, so mock interviews with recording and listening back to them and correcting yourself would really helpful in making sure you use your 45 mins wisely.

10. Last but not least it is improtant to take the things Simple, simple doesn't mean to be easy but Simple means to be bravity, consise. So try to habituate solving problems with consise and simple nature, I know its not easy I find that is the area which I lack and miss the opportunity.

Preparation

• I started in October 2020 with Leetcode premium, I have solved 360 questions by the time I finish my last interview with Google in April 1st week, after that I haven't given any problem solving interview.
  Leetcode progress
  

• I started with fundamentals from CTCI, MIT 6.006 and few other resources which I found from Leetcode and medium blogs. I've written first time Binary search code in October

• It's all with Practice, dedication, consistency and confidence, I would say I seen in one of the comment from lee215 you really need 99% sweat 1% talent.

• I started with daily problems on monthly challenge, I now participate every weekely contest I feel these challenges introduce many variety of questions to make us improbe Algorithmic thinking.

• Feel free to ask questions in the comment regarding preparation I would be happy to update post with what ever interest and help community.

Wish you all the good luck for your journey.

~
Thank You!