Tip 1 :Prepare resume really well
Tip 2 : Revise core data science, ML/DL concepts thoroughly
Tip 3 : Have good data science and ML projects in resume which you can explain clearly
Tip 1: Don't bluff/fake anything
Tip 2: Add good ML based projects that you can explain
This round was focused on breadth of topics ranging from Machine Learning, Deep Learning topics to MLOps and development topics
How do you deal with unbalanced data
How SVM hyperplane works
Roles of Generator in GANs etc.
Why Normalization should be done and when to not
When to use SQL and NoSQL type databases
Tip 1: Have a solid understanding of DBMS and SQL/NoSQL Databases
Describe how will you design a recommendation and search engine using ML system design
Tip 1: Study about docker, container, virtualization and some about AWS and cloud services
Tip 2: Have good understanding of system design
It was a full data structures and algorithmic round


Input: [[2,1,1],[1,1,0],[0,0,0]]
Output: 2
At T=0, only orange at (0,0) is rotten.
At T=1, oranges at (0,0),(0,1) and (1,0) are rotten.
At T=2, oranges at (0,0),(0,1),(1,0),(0,2) and (1,1) are rotten.
No fresh oranges are left after T=2.
Use multi-source BFS on grid



Can you solve each query in O(logN) ?
To Find Peak Element in rotated array
We can return the mid, if we found the peak in the mid i.e nums[mid-1]nums[mid+1] else we'll check if the peak can be found in right subarray or left subarray.
If the left subarray is unsorted that means peak is in the left subarray i.e if nums[low]>nums[mid]
else the peak is in the right subarray.
To find the target after peak (k)
Once the peak is found i.e k then we can easily search for elements with updated mid index as
(mid+k+1)%n , where k is the peak or point from where array is rotated.

Here's your problem of the day
Solving this problem will increase your chance to get selected in this company
What is recursion?