Please leave Uber engineer bashing on those other threads. I'm asking about the business model and don't have any beef with any engineers anywhere. That said, I don't understand what Uber needs 3000+ engineers for. Even post-layoff org size seems way too big for what the business needs and can support. The main app and technology haven't changed much over the years. I get that they do more than just ride-sharing now (eats, delivery, etc.). They're also scaling their services and investing heavily in autonomous vehicles, but even so, I don't get what hard problems they're working on that require such an army of high-paid engineers. Seems like a rival could compete with them easily by offering basic ride hailing with cheaper fares and higher fare share percentage for drivers, and just spend much less on R&D (in fact, there's a ride share in Austin that does this). The only problem with Uber's profitability isn't that mobile ride-hailing isn't profitable. It's just that the engineering team is just way too big. Is there something I'm missing?
What’s your background OP?
5 years at msft in Windows (~5000 engineers at the time), 1 year at a 1 dev startup, 4 years at Tableau (~1100 devs, 400 when I started). I've seen different size teams and what they can do, I just think I'm missing something about the business.
Amazon wants to know if you're an engineer yourself.
Freight (potentially a money maker), flying saucers, rewards program, map and routing optimization, various applied science and machine learning platforms, finance and payment systems, image processing, etc.
Do you need 600 devs for each of those?
Me - no, a two pizza team for each :-) But then becoming a VP or a Director is not going to that easy.
While there are duplication works at Uber due to incompetence, the actual engineering work involved in running Uber is nothing to laugh at. I will give you an example. How hard do you think it is to figure out a rider's location accurately and pick him up properly?
Yeah that's probably a non trivial problem, but two counterpoints: 1) how often does Uber update that algorithm? Seems like pickup experience is about the same over time. 2) Ride does a reasonable job at pinpointing rider location and crunchbase says they have 11-50 employees total.
Are you really an engineer? I am sorry if you are but your questions/counterpoints seem very odd. Very non-tech PM-ish.
It’s not just that. It’s matching rider to driver and doing it efficiently. Eats is even more complex, you need to match driver to consumer to restaurant and do it efficiently. Also lots of ML works goes into pricing for riders/drivers. Also there is the infra org which provides the platform for running all this.. I would actually be interested to see how many engineers lyft has for running their core business.
I read an article that Lyft makes less money per ride than uber because Lyft ends up paying more infra costs (mostly to AWS). This is one of the reasons uber builds everything in-house. Most vendors have margins. However, with their reputation and stock situation, I don’t think they can hire good engineers anymore. So it may become harder to manage these infra efficiently. Twitter experienced this few years back and we have been working on going to hybrid model
@twitter I doubt that article took into account the r&d cost for all those engineers who built and maintain it. Probably just compared the raw server costs which looks better for Uber. With r&d included wouldn't be surprised if they were equal or favorable to Lyft.
Take a look at the last earnings report. If you can explain to me a way to turn a profit by having a smaller eng team, I’ll personally refer you for the CFO position.
Engineer is a scaling asset. At smaller engineering team size, each engineer contributes a big portion to your product. This comes with diminishing returns, each engineer you add will add less and less to your product. But when you are a company that’s this big, a 0.1% increase to your business revenue from an additional engineer will still be worth way more than what you pay him for. Also, there are probably less than 2-300 mobile engineers working on the app itself. You worked on windows and tableau, both are pretty frontend heavy, with pretty low requirements for backend scaling, and you had 5000 and 1000+ engineers. Why would you think 300+ mobile engineers building Uber’s front end is too many? Uber has a backend that’s a lot more complicated than either tablet or older windows systems which is why we have a lot more backend engineers. Also, as someone mentioned, we have our own stack and don’t use aws. That easily saves us 1-2000 engineer’s worth of pay already.
1-2000 is a stretch. Recruiting, salary, office space, benefits, employment taxes, administration, etc. will bring the average engineers to let's say $300,000. So your figure becomes 300 million - 600 million per year. How many datacenters do you need to run again, factoring in availability and latency requirements? Whats your power bill per datacenter? What utilization do you get out of your hardware? If you're on AWS and I give you 1-2,000 engineers and say make as much money as possible would you start a project to migrate off AWS to your own infrastructure?
Icebergs also float.
Icebergs float because ice's mass per volume is less than that of water. Uber floats because it is propped up by outside funding for now. It's a totally different thing. For Uber to become an iceberg it needs to make more money per ride than it spends per ride. For that to happen it needs to charge more per ride (probably not possible), share less with drivers (probably not possible) or spend less on engineering. Or get driverless working. Seems like investors are essentially taking a moonshot on driverless rides, and if that doesn't pan out, the only other option is to decrease per-ride r&d overhead.
Over the years I have learned that engineering size does help the business even though a naive engineer things the problem are all solved
Yes, engineering size helps. But does it help enough to justify the massive cost? That's the big question.
India
Yesterday
669
'Hindutva': The Radical Hindu Ideology That Seeks to 'Push Christianity Out of India’
Personal Finance
Yesterday
1237
Thank you AAPL and NVDA
Tech Industry
Yesterday
1692
Do people underestimate E6 role at meta?
World Conflicts
Yesterday
535
Is "From the River to the Sea" So Wrong?
Tech Industry
Yesterday
414
How much should your net worth be by age 30?
And here I am, impressed with just 3000 engineers running Uber.