As of today, I've been back at the 'Soft for exactly a year. I was reminded of this by an email from the HR folks congratulating me on my "first year at Microsoft" and telling me that now would be a good time to talk to my manager about my career interests and aspirations. Riiight.
In any case, I thought this pseudo-milestone merited a quick retrospective on how the year has gone, work-wise. So, in no particular order, some thoughts:
- C#, and the Visual Studio IDE, are pretty damn cool. No more worrying about leaking or corrupting memory [well, most of the time at least], auto-completion provided by Intellisense, a powerful set of class libraries -- it's enough to make me only use Emacs [my previous editor of choice] and C/C++ when I absolutely have to.
- I've learned about, and started to build, some very interesting distributed systems stuff: load management via SEDA-like architectures, highly-resilient Internet services based on Recovery-Oriented Computing principles, applying machine learning to monitoring and failure detection in distributed systems, and using data stream mining to understand service performance. [Large-scale Internet services have a ton of interesting research problems, which is why papers like the Dynamo, GFS and Chubby papers have been accepted in top-tier conferences like SOSP and OSDI.]
- The team I'm on is absolutely top-notch. Razor-sharp folks, without the usual attendant egos, with everybody just focused on building the best system possible. It's also helped that we haven't had to deal with any political BS, due to the great fire cover provided by management.
- Being a manager is a lot easier when you a) have team members that need very little in the way of "management" and b) don't have to try to sell them something you don't believe in yourself.
- I've said it before, and I'll say it again: I have no regrets about leaving grad school. I'm much more an engineer than a scientist, in that I'd much rather build something than try to figure out how it works, and [synthetic] biology simply isn't at the point yet where you can build interesting systems without doing a ton of basic research along the way. And since I suck at building physical artifacts, building digital ones is as good as it gets for me.
In summary, it's been a good year work-wise.
