Skip to main content

2 posts tagged with "HAMi"

View All Tags

HAMi-core Adopted by NVIDIA KAI Scheduler: GPU Sharing Enters the Hard-Isolation Era

· 10 min read
HAMi Community

The integration target here is strictly HAMi-core, not the full HAMi platform. KAI Scheduler keeps its own scheduling capability and brings in HAMi-core to provide GPU memory isolation.

In June 2026, two core PRs were officially merged into the NVIDIA KAI Scheduler main branch. This means HAMi's GPU memory hard isolation will ship as a built-in feature in the next KAI Scheduler release. Cloud-native GPU scheduling has officially moved from "cooperative sharing" into the "hard isolation" era.

Validating AI Agent-Driven GPU Management on Kubernetes with HAMi and kagent

· 6 min read

Source: mesutoezdil.substack.com
GitHub Repo: kagentWithHami
Chinese translation by Jimmy Song, originally published on WeChat


One physical NVIDIA L40S virtualized into 10 vGPUs with HAMi. An AI Agent deployed as a Kubernetes CRD via kagent. Agent-to-Agent delegation, GPU pod creation, overcommit protection - all driven by Llama 3.3 70B with no closed-source dependencies.

CNCFHAMi is a CNCF Sandbox project