If you’re working on an AI project, you know the feeling: it’s like ‘The Hunger Games’ out there for GPUs. Either they’re all booked, sitting idle somewhere you can’t use them, or you’re juggling AWS, GCP, and on-prem like a madman.
I’m sure I’m not the only one who’s experienced this frustration. So, how do you deal with GPU shortages or scheduling? Do you have a magic scheduler or is it just a matter of Slack messages and crossed fingers?
I’d love to hear your war stories and learn from your experiences. Do you have any tips or tricks to share for managing GPU resources? What tools or strategies have you found most effective?
Let’s commiserate and collaborate to find ways to overcome this common pain point in AI project management.