I applied through my university, and was recommended for data science summer internship at Microsoft's main campus. Phone screen: 1 hour phone call with a principal data scientist. Discussed my background, then was asked some questions about statistics/probability and also on machine learning algorithms. Felt pretty conversational and relaxed, although you were expected to know your stuff. On-site: 5 back to back interview sessions (1 hour each). First was discussing projects I worked on, and felt more behavioral (although technical questions were asked). Next three were technical sessions in experimental design/hypothesis testing, machine learning/data engineering with white board coding (SQL type rather than algorithms), and statistics/probability, in that order. Final interview was mix of behavioral/case study session with more business oriented. Many of the questions on-site were framed in case studies (A/B testing, defining and monitoring performance metrics, etc), and I was expected to think through the entire process. None of the technical questions were super difficult nor were there trick questions, but you should know your basics in stats/machine learning/programming to do well. Also, you should ask lots of good questions (always come prepared!) and show enthusiasm for the position.

Feeling thrilled to have accepted the offer, I look back on this interview process as both challenging and rewarding. It began with a technical round where I tackled a complex A/B testing design question, followed by a SQL query on user retention. The final stages included a system design discussion that had me sweating, but honestly, the system design section on PracHub prepared me well. Each step pushed my limits, but the culmination was definitely worth it, landing me a role at Microsoft.