Oh boy. Not even gonna bother with a throwaway, draw your own inferences about whether I'm taking the thread title too literally. My current responsibilities:
(1) Architect for significant updates to one of our financial projection models. The requested timeline for results from these changes has been changed (directly by C-suite) from mid-November to late October to ASAFP over the course of the last week. Said model was built by non-programmers using a 20-year-old modeling platform; the lead architect was a new grad with a CS minor. Ever tried debugging a runtime error in a >1000-line C++ function, with no error messages, using nothing but cout? Try it sometime, if you've got a good therapist.
(2) PM for an internal application that executes the projection model in (1) on our in-house grid computing platform. The developers and tester for this are great, which helps. But none of them had seen the codebase until a month ago, since the previous developers and tester all moved to different teams since the last time we worked on this application. They also don't fully understand what the application is doing (I haven't had time to get them totally up to speed, though I'm working on it!), which means I get to build a lot of the testing tools myself, the highlight of which has been writing a FoxPro DBF parser in R (long story).
(3) Reviewer for the twisted hellscape of Excel workbooks that makes up one of my old (~2 years ago) team's processes, since the guy who was supposed to be reviewing them quit out of the blue (can't say I blame him) and my old team apparently couldn't find anyone else to do it. The old team usually gives me a heads up when they make substantial changes to parts of the process while I'm reviewing them, so I've got that going for me, which is nice.
I should mention that I'm ostensibly a data scientist, but sometimes management seems to take the idea that data scientists should be generalists a little too literally.
(1) Architect for significant updates to one of our financial projection models. The requested timeline for results from these changes has been changed (directly by C-suite) from mid-November to late October to ASAFP over the course of the last week. Said model was built by non-programmers using a 20-year-old modeling platform; the lead architect was a new grad with a CS minor. Ever tried debugging a runtime error in a >1000-line C++ function, with no error messages, using nothing but cout? Try it sometime, if you've got a good therapist.
(2) PM for an internal application that executes the projection model in (1) on our in-house grid computing platform. The developers and tester for this are great, which helps. But none of them had seen the codebase until a month ago, since the previous developers and tester all moved to different teams since the last time we worked on this application. They also don't fully understand what the application is doing (I haven't had time to get them totally up to speed, though I'm working on it!), which means I get to build a lot of the testing tools myself, the highlight of which has been writing a FoxPro DBF parser in R (long story).
(3) Reviewer for the twisted hellscape of Excel workbooks that makes up one of my old (~2 years ago) team's processes, since the guy who was supposed to be reviewing them quit out of the blue (can't say I blame him) and my old team apparently couldn't find anyone else to do it. The old team usually gives me a heads up when they make substantial changes to parts of the process while I'm reviewing them, so I've got that going for me, which is nice.
I should mention that I'm ostensibly a data scientist, but sometimes management seems to take the idea that data scientists should be generalists a little too literally.