Level: Staff-Level
Round: Full Journey · Type: Multiple Types · Difficulty: 6/10 · Duration: 300 min · Interviewer: Unfriendly
Topics: Breadth-First Search, Depth-First Search, System Design, Behavioral Questions, Graph Algorithms, Topological Sort
Location: Seattle, WA, US
Interview date: 2025-12-20
Got offer: False
I interviewed for an MTS role at Microsoft after a recruiter reached out on LinkedIn.
Phone Screen: The interview started with a discussion about my projects. Then, I was asked a question involving a graph representing GPUs, where GPUs are linked and belong to nodes. The first question was to group all GPUs belonging to the same node. I solved this using BFS/DFS.
The follow-up question involved dividing the GPUs into a fixed number of X groups to maximize the number of links within each group. I discussed some heuristic approaches.
Virtual Onsite:
Background Deep Dive: I was asked to describe the overall architecture of a major project I worked on and answered some standard behavioral questions.
System Design: The system design question was to design YouTube live comments. I didn't explain well how to handle popular live streams and why Kafka was used instead of other message queues. Given my years of experience, I felt I performed satisfactorily.
Hiring Manager Chat: This round was mostly behavioral questions. The hiring manager seemed to be looking for someone more senior.
Coding: The coding question was on topological sort, which was not difficult. The follow-up was to output the topological sort order.