--- name: coreweave-sdk-patterns-1d88ea96cf88 description: 'Production-ready patterns for CoreWeave GPU workload management with kubectl and Python. Use when building inference clients, managing GPU deployments programmatically, or creating reusable CoreWeave deployment templates. Trigger with phrases like "coreweave patterns", "coreweave client", "coreweave Python", "coreweave deployment template". ' allowed-tools: Read, Write, Edit version: 1.11.0 license: MIT author: Jeremy Longshore tags: - saas - gpu-cloud - kubernetes - inference - coreweave compatibility: Designed for Claude Code --- # CoreWeave SDK Patterns > **Community-contributed.** Not affiliated with, endorsed by, or sponsored by CoreWeave, Inc. CoreWeave is a registered trademark of CoreWeave, Inc. ## Overview CoreWeave is Kubernetes-native -- use kubectl, Kubernetes Python client, or Helm for programmatic management. These patterns cover GPU-aware deployment templates, inference client wrappers, and node affinity configurations. ## Instructions ### GPU Affinity Helper ```python # coreweave_helpers.py from dataclasses import dataclass @dataclass class GPUConfig: gpu_class: str # A100_PCIE_80GB, H100_SXM5, L40, etc. gpu_count: int = 1 memory_gb: int = 32 cpu_cores: int = 4 GPU_CATALOG = { "a100-80gb": GPUConfig("A100_PCIE_80GB", memory_gb=48, cpu_cores=8), "h100-80gb": GPUConfig("H100_SXM5", memory_gb=64, cpu_cores=12), "l40": GPUConfig("L40", memory_gb=24, cpu_cores=4), "a100-8x": GPUConfig("A100_NVLINK_A100_SXM4_80GB", gpu_count=8, memory_gb=256, cpu_cores=64), } def gpu_affinity_block(gpu_class: str) -> dict: return { "nodeAffinity": { "requiredDuringSchedulingIgnoredDuringExecution": { "nodeSelectorTerms": [{ "matchExpressions": [{ "key": "gpu.nvidia.com/class", "operator": "In", "values": [gpu_class], }] }] } } } def gpu_resources(config: GPUConfig) -> dict: return { "limits": { "nvidia.com/gpu": str(config.gpu_count), "memory": f"{config.memory_gb}Gi", "cpu": str(config.cpu_cores), }, "requests": { "nvidia.com/gpu": str(config.gpu_count), "memory": f"{config.memory_gb // 2}Gi", "cpu": str(config.cpu_cores // 2), }, } ``` ### Inference Client Wrapper ```python # inference_client.py import requests from typing import Optional class CoreWeaveInferenceClient: def __init__(self, endpoint: str, timeout: int = 30): self.endpoint = endpoint.rstrip("/") self.timeout = timeout self.session = requests.Session() def generate(self, prompt: str, max_tokens: int = 256, **kwargs) -> str: resp = self.session.post( f"{self.endpoint}/v1/completions", json={"prompt": prompt, "max_tokens": max_tokens, **kwargs}, timeout=self.timeout, ) resp.raise_for_status() return resp.json()["choices"][0]["text"] def chat(self, messages: list[dict], **kwargs) -> str: resp = self.session.post( f"{self.endpoint}/v1/chat/completions", json={"messages": messages, **kwargs}, timeout=self.timeout, ) resp.raise_for_status() return resp.json()["choices"][0]["message"]["content"] def health(self) -> bool: try: resp = self.session.get(f"{self.endpoint}/health", timeout=5) return resp.status_code == 200 except Exception: return False ``` ### Deployment Template Generator ```python import yaml def generate_inference_deployment( name: str, image: str, gpu_type: str = "a100-80gb", replicas: int = 1, port: int = 8000, ) -> str: config = GPU_CATALOG[gpu_type] return yaml.dump({ "apiVersion": "apps/v1", "kind": "Deployment", "metadata": {"name": name}, "spec": { "replicas": replicas, "selector": {"matchLabels": {"app": name}}, "template": { "metadata": {"labels": {"app": name}}, "spec": { "containers": [{ "name": name, "image": image, "ports": [{"containerPort": port}], "resources": gpu_resources(config), }], "affinity": gpu_affinity_block(config.gpu_class), }, }, }, }) ``` ## Error Handling | Error | Cause | Solution | |-------|-------|----------| | GPU class not found | Typo in node label | Use exact values from `gpu.nvidia.com/class` | | OOM on inference | Model too large for GPU | Use larger GPU or quantized model | | Connection refused | Service not ready | Check pod readiness probe | ## Prerequisites - A namespace-scoped Kubernetes credential and endpoint from the approved environment. - An image, GPU class, and resource budget reviewed for the target workload. - A secret-manager reference for private registry or model access; never pass tokens into generated YAML or application logs. ## Output - A reusable client or deployment manifest pattern with explicit GPU resources and affinity constraints. - A readiness-aware request path that distinguishes unavailable services from a valid application response. - A generated manifest that can be reviewed, versioned, and rolled back before apply. ## Examples Generate a manifest, inspect it for the expected namespace and GPU resource limit, then apply it first in staging: ```python manifest = generate_inference_deployment('summarizer', 'registry.example/summarizer:v1') open('summarizer.yaml', 'w').write(manifest) ``` ```bash kubectl -n inference-staging apply --dry-run=server -f summarizer.yaml kubectl -n inference-staging apply -f summarizer.yaml kubectl -n inference-staging rollout status deployment/summarizer --timeout=10m ``` If validation or rollout fails, retain the reviewed manifest and redacted events; do not broaden the client credential or bypass the admission policy. ## Resources - [CoreWeave GPU Instances](https://docs.coreweave.com/docs/platform/instances/gpu-instances) - [Kubernetes Python Client](https://github.com/kubernetes-client/python) ## Next Steps Apply patterns in `coreweave-core-workflow-a` for KServe inference deployments. --- Source attribution: jeremylongshore Original project: https://github.com/jeremylongshore/claude-code-plugins-plus-skills/blob/ab85a061f7bcc039dac13ee193de7a4cd4388571/plugins/saas-packs/coreweave-pack/skills/coreweave-sdk-patterns/SKILL.md License: MIT License information: https://opensource.org/license/mit Original upstream instructions; downloaded from Mega AI Hub. ## Upstream license notice MIT License Copyright (c) 2025 Jeremy Longshore Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. # Mega AI Hub resource metadata Creator / publisher: jeremylongshore Original project: https://github.com/jeremylongshore/claude-code-plugins-plus-skills/blob/ab85a061f7bcc039dac13ee193de7a4cd4388571/plugins/saas-packs/coreweave-pack/skills/coreweave-sdk-patterns/SKILL.md Attribution basis: Source repository record Fields of work: Development & Engineering Roles: Backend Developer, DevOps / SRE, ML / AI Engineer Specialties: API, Backend Tools: Kubernetes, Coreweave Country applicability: Country not specified Classification basis: Source catalog labels Instruction SHA-256: a8a0970476ea361963d1206d5d605bc7780215474bdf23b4b1d7f084b082fded