The initial Teams video pre-screen involved whiteboarding a theoretical design, followed by a six-hour interview circuit. Four one-hour interview sessions took place on a single day. Each interviewer decided whether I would be writing code or whiteboarding the solution, though sometimes I was given the flexibility to choose. Two interviewers were clear with their requirements and offered helpful hints. The other two, however, did not present the requirements clearly. As always, if I had seen the particular LeetCode questions before the interview, I would have performed much better, partially because I wouldn't have needed to rely on the interviewers for clarification.

All questions were what I would classify as "LeetCode medium" difficulty.

I feel that participating in these interviews with Microsoft, Amazon, and other companies is a decent way to prepare. I haven't really encountered technical interviews harder than those from Microsoft and Amazon. However, one must truly grind for several months (after work) or a few weeks (if not holding down a full-time job) to encounter enough of these potential LeetCode problems to have ideally seen them all beforehand. This has always been my gripe with these interviews: anyone who passes them has likely solved the same basic problem on HackerRank. It's not truly a test of coding ability so much as it is a test of memory.

Each interviewer is supposed to cover one or two of Microsoft's core values. They ask a question to give you the opportunity to describe something about your work history that demonstrates the core value(s).

In practice, only around half asked these kinds of non-technical questions.