HAMi Lab 17: a verified 60-minute guide from KitOps ModelKit to SGLang inference

HowDevelop · x · 2026-09-27

HAMi's Lab 17 is a verified 60-minute tutorial showing how to package a model as a KitOps ModelKit (versioned OCI artifact on Jozu Hub), pull and unpack it in a Pod via a custom kitunpacker initContainer, schedule it with HAMi's gpu/gpumem/gpucores parameters, and serve it with SGLang behind an OpenAI-compatible API — optionally co-locating a vLLM Pod on the same physical GPU. Includes full architecture diagrams and prerequisites (kind + H100 cluster).

Related event: KitOps ModelKit and HAMi Enable End-to-End GPU Inference Pipeline(2 posts)→

Original post →

More from coding & agent

coding & agent channel →