Haven't heard of XNA in years...

Try hanging out in the Mobile and Console area.

XNA is the only sanctioned tool for hobby X360 development. With an install base of over 75 million, it is still quite relevant even if MSFT is scaling back services for the older console.

True XNA hasn't changed for years, but neither has the X360 hardware. It was a cross-platform library for DX9-era cards on PC, X360, and certain phones. The library and the API can remain fixed when the hardware is also a fixed target.

Strangely enough, our mobile teams have mostly all switched to Unity over 2 years ago, and whatever work we do on MS consoles doesn't use XNA.

Possibly not the norm though.

Are your mobile teams doing hobby development, or are they a licensed developer?

I'm guessing the latter, in which case a major engine is likely the best choice.

Bit of both actually (as strangely as that might sound).

That being said, I fear we're derailing this topic, so my apologies for this tangent.

Oh yeah, sorry.

What do you guys think what is it about such a portfolio that could convince a AAA (or not) developer to hire a person? Note here I am only talking about the portfolio not the social skills. Let's say I want to apply for a 3D-programmer job.

The things on the list are good, but really the employer will likely only care if you appear to be able to do the job they need done.

If your portfolio shows you are specialized only in writing shadow volumes and similar code, that won't help much if they're looking for someone to improve their terrain system or someone to work on network code. Similarly if your portfolio shows you specialize in networking and communications it won't help much if the employer is looking for gameplay programmers. Put in stuff that showcases what you have done and want to do more of.

The point of a portfolio is do provide evidence that you can do the job.

Show that you can make games by actually making games.

If you have a few years of experience in the industry you can just say "I did such-and-such on each of these games." That is strong evidence that you can do the job of making games.

If you haven't done the job before then you need to find another way to demonstrate you can do it. If your portfolio is a bunch of half-finished school projects with source that won't compile and have major structural errors, that is bad evidence. If your portfolio is a completed 3D network game that includes both an executable game that is fun to play and good-looking source, that is good evidence.