Working in astronomy, especially in a big collaboration, is not just about working on your own little project over in a corner. A good supervisor will push you to get involved with other collaborators and to take on responsibilities from other members (who, due to the transient nature of post-doctoral research contracts, will soon leave).
This can lead to the very strange scenario that you're very junior and yet the only person who can do a certain task. In my case, it's finding the galaxies in our images (or, in astro-speak, "producing the catalogues/cats for the maps"). We had a big team working for a while to produce the software to find the galaxies, but a few years later they've moved on to other projects. So I, like many other PhD students, are "standing on the shoulders of giants" to quote Newton (in a non-sarcastic way!).
This is great when everything's going smoothly. But, when things go wrong (like today, when we've spotted an error in the data release!) it can mean digging through a lot of code you didn't write (that is well documented, thankfully) to make code that ran three years ago work. Really not sure why there was a problem in the first place but hey-ho, I can address that when the code is running again...