I interviewed as a senior in college. The interviewing process consisted of two phases for me: (1) a single on-campus interview when Microsoft visited my college campus, and (2) a day-long on-site interview at Microsoft's campus in Redmond.

The on-campus interview was with an engineer and consisted of some simple technical questions. We also talked a bit about some programming projects I had worked on at school. I believe they were looking for basic technical chops and enthusiasm for engineering.

The on-site interview consisted of one interview with an HR person and 4 interviews with engineers (two from one product/team, and two from a completely different product/team). The interview with the HR person was non-technical and focused on why I was interested in becoming an engineer. The interviews with engineers were all technical in nature. They mostly consisted of programming and algorithms questions of moderate difficulty --- the usual affair. I also had lunch at some point with one of the engineers I ended up interviewing with.

One of the technical interviews was fairly nonstandard. I was asked to come up with a design (hardware/software) for a robotic telescope. The question was very open-ended, though the interviewer of course gave some guidance. I believe the purpose of the question was to gauge the ability of the candidate to design a (real-world) system.

Describe how to design a robotic telescope. Talk about hardware (robotics/sensors); talk about software (linear algebra/control).