Staff platform engineer

Own the compute and data planes underneath every stack we ship — the layer clients never see and can never afford to have fail.

You'll work on the substrate

multi-region compute scheduling, autoscaling under bursty inference load, and the storage and networking seams that everything above depends on. This is a small team with real ownership. You will not be handed a ticket queue — you'll be handed a constraint and trusted to design your way out of it.