The interview involved three 45-minute, back-to-back technical phone interviews. They asked me to solve and implement solutions to basic algorithmic problems, of the type you would encounter on TopCoder or similar platforms. The questions weren't easy, but they required technical maturity and confidence with these kinds of problems.
You could choose your preferred programming language. However, one of the interviewers assumed I would use Java, while I had decided on Python. Therefore, I suggest having one preferred and one backup language; the backup should be either Java or C++.
During each interview, you may solve more than one problem. Sometimes these problems are incremental (building on solutions to previous ones), and sometimes they are brand new.
After having taken these interviews, I feel that spending some time on TopCoder and getting acquainted with those problems is worth the time.
The following metrics were computed from 2 interview experiences for the Google PhD Software Engineering role in United States.
Google's interview process for their PhD Software Engineering roles in the United States is extremely selective, failing the vast majority of engineers.
Candidates reported having very negative feelings for Google's PhD Software Engineering interview process in United States.