What are some pros and cons of Uber vs Lyft from data science perspective? I am primarily interested in comparison on three parameters: 1) technology stack 2) compensation 3) work-life balance Thanks
DS at Uber, so obviously biased One way to view the data situation is that Lyft is like Uber a few years ago. Really only 1 product in 1 country, which could get boring after a while. However, you likely have much more responsibility since it's a leaner team and it might feel like a scrappier startup. Uber objectively wins on the data side. We have so much more and it spans the globe with a handful of products. We also have an enormous data infra team maintaining it. Tech stack at Uber for DS is: data backend is Vertica, Hive, and Presto. EDA is usually done in R markdown or IPython notebooks and production is all in Python with a little bit of Go. We also just got a crop of GPUs for deep learning. Work life balance depends on your team. I'd reckon since Lyft is losing their only market right now, it's probably not the place to rest and vest. They probably work extremely hard in order to win market share and beat Uber to new features. The "always on" work all the time culture has definitely chilled out since Travis left. Uber just rehauled comp so it's in-line with market and maybe a tad higher. Equity is probably so-so these days but it's lower risk that Lyft. I also don't believe Lyft is still giving out equity at least as of 8 months ago.
Lyft has ownership! Uber has more data but also an oversupply of data scientists
Oversupply? With mapping, SDCs, a more advanced marketplace, and more products/services (like trucking and eats), we have so many more DS problems we're actively working on. Every DS team I know about has headcount to hire more.
Uber is much better with more data, more infra, deeper talent pool. And Uber has more products more markets , more ways to apply your knowledge. Work life balance is pretty good now, they have consciously changed for the better.
I do wonder how much ownership you'd get at Uber vs Lyft. How often can you lead an initiative or project? Especially if you have your own ideas, as I assume the OP does with his interest in transportation analytics.
I work on a relatively large team. ~15 data scientists and have a ton of ownership. I'm able to pick and choose interesting projects within the scope of our team/OKRs and run with it. I can't speak for other teams, but there are definitely places here that give you a lot of responsibility.
I think Uber gives a lot of freedom and ownership, and opportunities for impact. That's been my experience.
At Lyft, you have HUGE scope. In all of marketplace we have 10 data scientists
It depends on what you want to do. Uber has more data and it is in more countries, so if you want to do some cutting edge application with unique data and different kinds of regulation, Uber is your place. Lyft has a smaller team, so you'd probably have more ownership and you can be main responsible for a product. Both are gratifying paths and have smart people doing good work. You get too choose, but don't drink too much cool aid.
All good info here, its best to interview with actual team at both places to find personal fit instead of paralysis by analysis here.
Why the rideshares specifically?
I am interested in solving data science problems in transportation.
I don't know their stack but I've met people at both. Lyft probably has more pleasant culture and people. Work at Uber might be more cool, although Lyft is positioning itself well for the future. I wouldn't trust work life balance at Uber (What is this strange "work life balance" thing you speak of? You realize you're contributing to changing the world!) Uber has more data overall, and across the world. And they are more directly trying to develop both mapping apps and self driving cars, so they should run into interesting problems. I believe Uber's comp is above average although their stock options are less than dependable.