Microsoft SDE interview | 2 years experience

I am currently working as a Senior Software Engineer with Samsung RnD, Bangalore, having completed my Bachelor's in Computer Science. I applied for an SDE 1 position at Microsoft's Hyderabad office.

The interview process involved three distinct rounds:

Round 1: Techno-Behavioral

This round began with standard behavioral questions. The interviewer asked me to introduce myself and discuss the projects I've been involved with, expecting detailed explanations from my resume. They also inquired about my team size and the number of people on my project. A key behavioral question involved describing a technically challenging situation I faced, how I overcame it, and a follow-up on how the solution was accepted, particularly focusing on developing a cleaner solution after the initial deadline.

For the technical part, the interviewer explicitly stated they valued the thought process over just the solution and that the problem would be ambiguous, requiring me to ask clarifying questions. The problem given was: Given a list of integers represented as strings, remove every Kth element and every element that is equal to K. I was expected to write test cases covering all corner cases, explain my expected output, and write modular, industry-standard code.

The round concluded with an opportunity for me to ask questions and highlight any significant achievements.

Round 2: Technical

This round started briefly with 'tell me about yourself' and a question about my alma mater. The main focus was on a data structures and algorithms problem: Given employee in and out times, determine the maximum number of employees present in the building at any given time. I was challenged to optimize the solution, and I proposed and coded an O(1) space and O(nlogn) time complexity approach.

I was again given a chance to ask questions at the end.

Round 3: Techno-Behavioral

The final round also began with 'tell me about yourself' and a question about my graduation. The interviewer then asked about my motivations for seeking a switch.

The core technical question in this round was more conceptual and leaned towards system design principles: Imagine a 'chunker' over a network that can split a given string into any order of chunks (e.g., 'This is a test string' could become 'This is', 'string', 'test string', etc.). I needed to determine what extra information each packet (or chunk) would require to reconstruct the original string accurately on the receiving end. My proposed solution involved sending the original string's total length and the chunk's starting index within the original string with every packet (e.g., '20|15|string'). On the reconstruction side, I described how I would initialize a string of the full length and use a list of yetToBeFilledIndexes to track and fill parts of the string as chunks arrived.