SWE at Microsoft

How it went

What they would tell you

How to prepare

Their background

Round by round

1. 1

   DSA Round 160 min

   First Problem: Max Consecutive Ones II used a sliding window with at most one zero flip, updating max length dynamically Second Problem: Rotten Oranges implemented BFS level-wise traversal to simulate the spread of rot across the grid

   • Max Consecutive Ones II
   • Rotten Oranges
2. 2

   DSA Round 260 min

   In my second-round interview, I was given the problem “All Paths from Source Lead to Destination”, with an extra task to print one of the shortest paths. I solved it using a Depth First Search approach and kept track of the paths as I explored them. While doing this, I also stored the shortest path found so far. To make the solution faster, I avoided checking paths that clearly wouldn’t lead to the destination. In the end, I explained the time complexity and how skipping unnecessary paths made the solution more efficient.

   • All Paths from Source Lead to Destination
3. 3

   High-Level Design Round60 min

   Asked to design new feature for music streaming service 'Top K Songs in the Last 7 Days' My approach was to use 7-day sliding window to keep track of how often each song was played each day. To efficiently find the top K songs, I maintained a min-heap of size K that dynamically updated as new data came in. Each day, the system would add the latest day’s data, remove the oldest, update the counts, and return the top K songs. For the system design, I included components such as a database for storing song data, a cache for quick access to song counts, a playback server, a top K songs server, an API gateway, and a load balancer. Overall, this round didn’t go as well as I hoped.

What came up