Add repo topic tags for discoverability

[resker]

Summary

Add topic tags to NVIDIA/topograph. The repo currently has no topics set — topic tags are the primary surface by which the repo is discovered via GitHub search and curated lists (e.g., topic:kubernetes).

Current state:

$ gh api repos/NVIDIA/topograph --jq .topics
null

Reference precedent — TRT-LLM has 5 topic tags (blackwell, cuda, llm-serving, moe, pytorch). For topograph, proposed set reflecting primary users, integrations, and ecosystem context:

kubernetes
slurm
slinky
topology
infiniband
nvlink
gpu
network-topology
topology-aware
scheduling
cncf
dra

Rationale for each:

Tag	Why
kubernetes	K8s is the primary engine target
slurm	Slurm is the other primary engine target
slinky	Slinky is the other primary engine target
topology / network-topology / topology-aware	What the project discovers
infiniband, nvlink, gpu	Fabric and accelerator types surfaced
scheduling	Downstream consumer concern (KAI, Kueue, etc.)
cncf	Integrates with KAI Scheduler (CNCF Sandbox)
dra	NVIDIA DRA is a first-class provider

How

Requires admin on the repo. Command:

gh repo edit NVIDIA/topograph \
  --add-topic kubernetes \
  --add-topic slurm \
  --add-topic slinky \
  --add-topic topology \
  --add-topic infiniband \
  --add-topic nvlink \
  --add-topic gpu \
  --add-topic network-topology \
  --add-topic topology-aware \
  --add-topic scheduling \
  --add-topic cncf \
  --add-topic dra

Or via GitHub UI: repo page → gear icon next to "About" → Topics field.

Reference

• Example: NVIDIA/TensorRT-LLM topics
• GitHub docs: Classifying your repository with topics