6 ms·

I thought game dev salaries were bad, but people still applied because of the thrill.

by fruit2020 4y ago

KyeRussell 4y ago
Those would be the roles that “everyone wants” as alluded to.
theshrike79 4y ago
In a team of 20 making a mobile game, you have 17-18 people who don't do backend and 2 people who do. Those two are the ones that are invaluable, there's a literal queue out the door to fill the sexier positions of game designer/producer/2d/3d artist/unity developer and whatever. You can just pretty much plug in a new Unity developer to the team, since the skill set isn't that hard to learn. ...but finding someone who can do 1M+ DAU game backend that's reliable and cost-efficient isn't something you just hire off the street.
- lylejantzi3rd 4y ago
  > but finding someone who can do 1M+ DAU game backend that's reliable and cost-efficient isn't something you just hire off the street. So, how do you get that particular skill? I've done backend development, but nothing with that kind of scale.
  - vorpalhex 4y ago
    You can still design the systems and try to test them. Build programs and throw load tests at them. Scaling software is more about careful choices than some ultra deep understanding (though really understanding your data source is very important).
  - bcrosby95 4y ago
    The difficulty really depends on the game. Some of these things are glorified webapps and 1M+ DAU is pretty trivial on the cheap. But if you have more real-time multiplayer interaction, it would take a lot more work.
  - theshrike79 4y ago
    Basically you're paying for someone who has seen shit fail 42 different ways and has a pretty good feeling/knowledge what works in which case based on their experience. And they also know which solutions are overkill. Simpler is always better, don't do stuff for your CV, build it so that you don't need to wake up at 3am on Christmas to fix some obscure deadlock issue. Log stuff so that you can tell most of the time what went wrong just by looking at the logs and graphs. But also be mindful of costs, logging everything to Cloudwatch can get surprisingly expensive. Benchmarking is important, you can use the logs to see where you're actually spending time and optimise it if possible. Sometimes the illusion of real-time is better than being real time. Everything that can be cached, should be cached. Utilise the TTL in Redis to make stuff simpler.
    - lylejantzi3rd 4y ago
      Ok, so how do I convince a prospective employer that I'm capable of these things? Well, except for seeing shit fail 42 different ways. We have a bit of a chicken and egg problem there. How can I have seen shit fail 42 different ways if I need to see shit fail 42 different ways to get the job?
      - theshrike79 4y ago
        Not all games are 1M DAU, you can leverage knowledge of smaller games first. Only the biggest games with huge marketing budgets go from zero to 1M DAU in a day, so there's always the possibility of slowly improving the system while the player count grows.

I thought game dev salaries were bad, but people still applied because of the thrill.

Those would be the roles that “everyone wants” as alluded to.

In a team of 20 making a mobile game, you have 17-18 people who don't do backend and 2 people who do. Those two are the ones that are invaluable, there's a literal queue out the door to fill the sexier positions of game designer/producer/2d/3d artist/unity developer and whatever. You can just pretty much plug in a new Unity developer to the team, since the skill set isn't that hard to learn. ...but finding someone who can do 1M+ DAU game backend that's reliable and cost-efficient isn't something you just hire off the street.

lylejantzi3rd 4y ago
> but finding someone who can do 1M+ DAU game backend that's reliable and cost-efficient isn't something you just hire off the street. So, how do you get that particular skill? I've done backend development, but nothing with that kind of scale.
- vorpalhex 4y ago
  You can still design the systems and try to test them. Build programs and throw load tests at them. Scaling software is more about careful choices than some ultra deep understanding (though really understanding your data source is very important).
- bcrosby95 4y ago
  The difficulty really depends on the game. Some of these things are glorified webapps and 1M+ DAU is pretty trivial on the cheap. But if you have more real-time multiplayer interaction, it would take a lot more work.
- theshrike79 4y ago
  Basically you're paying for someone who has seen shit fail 42 different ways and has a pretty good feeling/knowledge what works in which case based on their experience. And they also know which solutions are overkill. Simpler is always better, don't do stuff for your CV, build it so that you don't need to wake up at 3am on Christmas to fix some obscure deadlock issue. Log stuff so that you can tell most of the time what went wrong just by looking at the logs and graphs. But also be mindful of costs, logging everything to Cloudwatch can get surprisingly expensive. Benchmarking is important, you can use the logs to see where you're actually spending time and optimise it if possible. Sometimes the illusion of real-time is better than being real time. Everything that can be cached, should be cached. Utilise the TTL in Redis to make stuff simpler.
  - lylejantzi3rd 4y ago
    Ok, so how do I convince a prospective employer that I'm capable of these things? Well, except for seeing shit fail 42 different ways. We have a bit of a chicken and egg problem there. How can I have seen shit fail 42 different ways if I need to see shit fail 42 different ways to get the job?
    - theshrike79 4y ago
      Not all games are 1M DAU, you can leverage knowledge of smaller games first. Only the biggest games with huge marketing budgets go from zero to 1M DAU in a day, so there's always the possibility of slowly improving the system while the player count grows.