I was interviewed by Microsoft's advertisement group for a data mining position for Bing.com during the annual meeting of my field.

I interviewed with two women at the conference and answered questions related to my research experiences, skill sets, and some brain teaser questions.

Then, I was offered an on-site interview three weeks later in Seattle.

The on-site interview involved meeting with nine members from two different groups for the role. All talks were one-to-one and each lasted for about an hour.

They seemed to have different assignments to measure different aspects of my skills, and they were all quite friendly, but the questions were challenging.

The questions they asked were mostly about basic statistical knowledge, which I "screwed up" because I had neglected them for too long.

Overall, they were a nice group, and be prepared for basic questions in statistics.

Explain heteroskedasticity to us, and specifically how to solve it.