Topics
Posts grouped by subject.
Schedulers and Kubernetes platform 8
How work gets queued, placed and delivered on shared clusters.
GPU and AI infrastructure 7
Autoscaling, monitoring, scheduling and topology for GPU fleets, tried on emulators first.
Bare metal and fleet operations 6
Bringing servers up and keeping them healthy: BMCs, provisioning, RAID, drift, inventory and hardening.
ML platform and automation 6
Serving a model, moving data to it, and automating the pieces.
Power and cooling 2
From the outlet back to the transformer, and what GPU racks change.