TC 180, 6yoe, Seattle. I went from Systems Engineer to SDE after 2 years. Then SDE I > SDE II after another 2.5 years. Been SDE II for about 1.5 years. That SysEng experience hurts me at Amazon (waiting in line for L6 behind those more senior SDE IIs), but doesn't prevent me from interviewing for, say, a Senior SDE position at Microsoft (equivalent to upper SDE II at Amazon). I switched teams to get some ML experience, but after a year of zero ML experience, I'm either going to try one more team at Amazon, or look externally for an SDE role for 230+ TC. I really want to get into it. Some teams at Amazon allow SDEs to actually work on ML problems, and some severely bifurcate the science and eng teams. Currently on the latter. I think it would be very valuable for my career in the long run to get some solid ML experience, but I also really want to stop taking such a fat opportunity loss. I feel that I'll have a better chance to get this type of position internally rather than externally. I've taken all the relevant internal ML courses. Recently tried applying to an external position that wasn't even advertised as requiring ML experience (although it was a service team for an ML) and I lost out to... someone with "significant ML experience in production". :|
In exactly the same boat, wondering the same thing
My partner is an SDE at Amazon... And he's having your same problem, the promise for ML work is never delivered...
I feel that. What sucks is that I've seen roles that are called "Machine Learning Engineer". I feel like it has to be delivered somewhere. Just wish I could land on a team like that.
I'm in a similar boat as well, I recently got offers from different companies for an ML team and a cloud team, both with similar pay. I think this question is really difficult to answer when you're approaching the decision by trying to predict whether companies will start looking for more and more ML engineers and scientists while there is already a huge number of people that can be a mediocre SDE. My friend suggested a different viewpoint: he was viewing this as a question whether you can be "the best" at ML compared to being the best SDE. It seemed clear then that it will probably be nearly impossible to be a strong ML engineer or scientist compared to a bunch of others that go into that field with a PhD in our age, whereas you most likely have a higher chance on being a rockstar engineer at this point. It comes down to what you want to do in your life, but the grass always seems greener on the other side :P Disclaimer: I am still split between choosing ML team or cloud team rn..
Re-reading your question, i think you can make a smoother transition to ML if you move to am ML team internally first, which is what I did and how I ended up receiving an offer from an ML team outside of Amazon.
ML is like big data, the bubble will fizzle
The thing about ML at this point is, it's mostly libraries, now. Not very many people making new tech in this space, and those that are, have PhDs. You can start out with pretrained models, train a bit on top of it for your custom thing, and you're good. Most of the work nowadays seems more doable for engineers. I guess I'm not missing out on that much.