Modal engineer built Gang Scheduler on K8s Reconciler model — it just worked
emilyzsh · x · 2026-10-04
Modal engineer qjoyliu shares that she and coworker Colin built the Gang Scheduler about a year ago by adopting the observe: S, plan: S -> P, apply: P -> S' model from the Kubernetes Reconciler — then left it alone and it kept working. She notes Modal prides itself on being "kube-but-better," but mature systems like Kubernetes still hold real design wisdom worth learning from, and teases upcoming work on clusters.
More from Infra
- Why AI labs won't push on-device models: cloud inference is their business — AccBalanced · 2026-10-04
- uv fork runs dev sessions ~33% faster with 40-60% less disk, author still unsure it's enough — mitsuhiko · 2026-10-04
- KohakuFA: Blackwell flash attention kernel fixes 10x-1000x gradient bugs in FA4/cuDNN — bdsqlsz · 2026-10-04
- RWKV-7 G1k ships: pure-RNN reasoning with no KV cache, 16M-state 13B model — cephaloform · 2026-10-04
- Running a 27B Model on 16GB VRAM: NInfer 4080 Hits 262 tok/s Decode — roofkid · 2026-10-04
- NeoCloud Summit 2026 lands in SF Oct 8, gathering the GPU-native cloud ecosystem — AccBalanced · 2026-10-04