It's been a whirlwind this past week. I may have to go a bit slower on the game development stuff now that my full-time job has started but let me report what I have done so far.
I managed to get nanite enabled, but I feel as though the Mac version of Unreal might be horribly bugged, because first I tried building the nanite, and it took a very long time, which I eventually had to force quit as it was using 120 GB of application memory. After relaunching the engine, my landscapes were bugged completely so I had to rebuild it from the raw file all over again. Which wasn't too much of a problem since I didn't edit it, really. Even after that, I tried simply saving and it had the same hanging issue. I am really not sure what the issue was. I tried restarting my machine and verifying the installation to see if that had an issues, but time will tell if I run into them again. It may have been a freak accident.
I am also dealing with performance issues in general. My other colleague working on this doesn't have an actual GPU for his laptop so we are considering just having a remote machine that's at one location that we can remote desktop into to develop the game. Containers and VCS are also naturally things we are exploring.
Finally, we are exploring our options as to how to import assets for the game. For the time being, we will probably be using public assets since we do not have an artist at the moment, but it's enough to get going. Our game will heavily involve reinforcement-learning AI to control certain aspects of the game, using computer vision, so it will be an interesting process for sure. Learning more about things like AirSim to help.
As for the job, nothing too interesting yet. Obviously there's some training, and since our team is brand new, the lead devs are still trying to figure out how to actually get started, and get the go-ahead from our higher-ups considering costs and all that. We don't have to have to scrap everything and start over so we're being very careful. A stitch in time saves nine.
Currently, we are exploring the existing Salesforce application to see how it's set up and looking out for potential pain points to see how we can immediately improve the experience of our firm's specialists. My own current task is looking around for APIs so that we can do some analysis in Python, and I was told to use Simple Salesforce, though apparently SalesforcePy exists. We also have data in Microsoft Sharepoint which is accessed through Microsoft Graph, but I don't need to look as deeply into that for now. The end goal is to use pm4py to run some specific process analysis on all that. I also am to learn how to use Pandas to perform some data analysis on whatever's in Salesforce.
I am also trying to make a habit of taking note of EVERYTHING that I come across. My boss told me about a neat way of organizing notes called "zettelkasten" that I am already using a ton. It basically creates a node graph for you based off the notes that you write in it, using links and all that, so the data is organized kinda like a "brain," with emergent patterns.
Overall, the job so far is pretty dry and "unsexy," as one of my bosses describes it, for now, but as soon as we can get into the meat of the AI stuff, it'll probably be a lot cooler and I hope I can report more about what I learn in that regard.
Stay tuned. I think I will start keeping this journal weekly on Fridays.
Top comments (0)