The difficulty the cloud-services-team would have in offering a devops/infrastructure track for newcomers would be making the intro tasks small and well defined enough and finding the resources for proper mentoring. Much of our devops backlog tends to be complicated projects or things needing deep research. I'll bring it up with the team however and see if we can come up with any ideas. We are definitely interested in being more involved in mentoring projects, but we have been discussing things like GSOC/Outreachy internships that are more targeted and focused.

There may be some opportunities coming up in the work that SRE will be doing to prepare for a Puppet4 upgrade, but it would really depend on mentor availability and the repeatability of tasks.

Note that here we are talking specifically about projects done by Computer Science / Engineering students in their last year, working 4-5 months on a final career project. These students might be newcomers to Wikimedia, but not to programming and perhaps familiar with several pieces of our infrastructure / stack.

In T170242#3426029, @Qgil wrote:

Note that here we are talking specifically about projects done by Computer Science / Engineering students in their last year, working 4-5 on a final career project. These students might be newcomers to Wikimedia, but not to programming and perhaps familiar with several pieces of our infrastructure / stack.

4-5 hours/days/weeks?

(edited)

4-5 months. Note that these are not full time internships as in GSoC/Outreachy. They would be still studying while competing these projects (@Mariusmm @Toniher please correct me if I'm wrong).

From what I understood from the different meetings last Friday, they consider TFG to imply around 300 hours of work. http://www.uab.cat/web/estudiar/llistat-de-graus/pla-d-estudis/treball-de-fi-de-grau/x-1345468416862.html?param1=1263367146646

300hrs is probably about 2/3 of the time estimate for a GSoC/Outreachy internship. The rule of thumb that I've heard before for sizing those projects is 3 months of new volunteer time is roughly equal to 3 weeks of mid-level engineer time. So for this I'd guess we would be shooting for something that feels like a 2 week/80 hour project. Based on past experience I would roughly estimate mentoring for that at 80-120hrs over the 4-5 month period.

Well, somehow it looks like this is not going to happen? Or am I being too pessimistic / premature?

In T170242#3425293, @bd808 wrote:

I'll bring it up with the team however and see if we can come up with any ideas. We are definitely interested in being more involved in mentoring projects, but we have been discussing things like GSOC/Outreachy internships that are more targeted and focused.

@bd808: Wondering if there was any outcome? If not and if the WMCS crew cannot cut down smaller tasks in a suitable project, I'd propose to decline this task.

In T170242#3617844, @Aklapper wrote:

@bd808: Wondering if there was any outcome? If not and if the WMCS crew cannot cut down smaller tasks in a suitable project, I'd propose to decline this task.

The cloud-services-team has interest, but currently does not have the capacity to do the task decomposition and mentoring that would be required to offer a standing Featured Project. We have posted an Outreachy project as an initial experiment in directed volunteer collaboration. I would certainly like to revisit the idea of a Featured Project in the future.

The initial subject here mentions SRE as well, but I don't see that they were ever looped into the discussion.