Several people have asked me about how I think SCOPE went for my team, and this is my attempt to answer that question.
I don't think we got as far as we could have technically, but some of that was out of our hands. We thought we were going to be able to develop some models, get data to check how they worked (or even to help us develop some of the models we worked on later), and iterate from there. Based on when we got the data, there wasn't really time for iteration to happen. We compared the data to the models we had and drew conclusions about how well the model was performing, which told us we had some models working very well, some models with all the right dependencies but not the right multiplicative factors, and some models not performing well at all. In the cases where the models weren't performing well, we did have time to look through our data sets and compare data with itself to determine which parameters matter most and what effect they have on important figures of merit. So in that sense, I think we did what we could.
I'm somewhat disappointed in my own technical work because I don't think it was particularly creative. I spent a long time reading about topics that I thought could help me approach a particular problem, and after realizing that I'd let myself get in a hole of reading papers and not making progress, I ended up going with a very simple model that we knew wouldn't quite work. (At the time, we did expect to have enough time to adjust it.) I wish I hadn't let myself get stuck and that I'd actually been able to develop something a little more sophisticated. I think I could have done better at capturing what was going on but wasn't creative enough or didn't put in enough effort to learn some of the more difficult concepts that could have been relevant.
I was also the team's product owner, which mostly meant I was in charge of communication with Boston Scientific. I was also in charge of our midterm and final reports, which wasn't necessarily a responsibility of the product owner, but it was convenient to have both jobs. I'm happy with the work I did in this area. I wasn't sure about being product owner, but most of my work week to week was keeping and sending minutes from the phone calls with our liaisons and emailing our liaisons if anything came up between phone calls. This happened more during certain parts of the project than others. Any trip we took to BoSci took a fair amount of organization, and from late March through late April it felt like I was constantly emailing people at Boston Scientific about data. I was skeptical at the beginning about most of the communication between the team and the company having to go through one person, but it definitely worked better and kept everyone on the same page. This is a role I would take again, which I wouldn't have guessed at the beginning of the project.
In terms of the reports, I'm pretty happy with how they came out, particularly the final report. We were doing data analysis until a couple of days before the report was due, so I'm pleased that we were able to draw and communicate conclusions about how well the models work and what future steps should be for the models that aren't currently accurate. Brian returned the first draft of our report with a note that said it was "really good," and I was inordinately pleased about that.
Overall, I'm not sure this was a good SCOPE project generally. Our main liaison was our user, and while that was convenient for being able to schedule co-designs, it's a kind of odd dynamic (and it's not a normal dynamic for Boston Scientific teams). It was also an exploratory project to help an exploratory division; we were developing a model that could help the people who develop new stents have to make and test fewer physical prototypes. If we had fully finished our work it could have been useful, and some of it is useful now, but we're not confident that anyone will carry the project forward and get it to a really good place. That's sad because it's really interesting, and I think understanding some of the behavior that we were really just starting to get to at the end with the data sets could be helpful.
That said, this was the best possible project that I could have been on in terms of my skill set and interests -- not just this year, but in the recent past of SCOPE. So I'm glad it happened, and I enjoyed this project. I just wish we'd been able to get a little more done.
No comments:
Post a Comment