A bunch of miscellaneous notes about this process, as I see it:
* I assume that most of the interactions with students will happen with the individual projects and on their comms, and up to the projects which students and proposals they particularly like for their own project.
* When it comes time to select the chosen students for the whole org, we can have a zoom meeting with reps from the projects to try to weigh the relative merits and decide how to allocate the number of funded students we are allocated among the projects and their favourite proposals.
* People will come along asking to be mentors. Mentors don't need to be TSC members or major committers. For example, for a color scientist/developer who has not previously contributed to OCIO but knows it well may still be a fine mentor for a student. We will also get unsolicited offers to be mentors from people who have no obvious connection to the projects nor obvious relevant expertise, and I suggest that we politely decline with a note that says we expect mentors to have a prior connection to the project.
* I think I forgot to mention, Google funds each student ($3000-$6000, depending on their geographic location), but they also pay the organization -- $500 per student, $500 total to send a student or mentor to our conference (e.g. our siggraph open source day), and $2200 to spend sending mentors to the "mentor summit" that happens in October at Google HQ). We can decide later how to spend the money, but since we are already a well funded organization, my recommendation would be to pour it all right back into student outreach and support.
* Please do recommend this program to students you know, former interns, whatever. Promote it on mail lists and corporate channels. There's nothing wrong with encouraging people we already know and think will be good summer students into the program. (They must be enrolled university students and are expected to work full time to get the stipend.)