Skill UI

Accelerates CPU-bound Python workflows by translating NumPy, pandas, ML, graph, image, simulation, and IO code into NVIDIA GPU kernels using CuPy, Numba CUDA, Warp, cuDF/cuML/cuGraph, KvikIO, cuSpatial, and RAFT for dramatic speedups.

Modal is a serverless cloud platform for running Python code on demand, including GPU-accelerated AI/ML workloads. Use for deploying models, serving web APIs, batch processing, and scaling Python applications to cloud containers with automatic scaling from zero to thousands of instances. Supports custom container images, persistent storage, and sub-second cold starts for low-latency inference.

Accelerate scientific Python code on NVIDIA hardware using CUDA and RAPIDS libraries. Optimizes NumPy, pandas, scikit-learn, and ML workloads for GPU execution. Includes profiling, memory management, and kernel optimization to ensure correctness and performance gains.

Automate the entire deployment lifecycle for ML inference services on CoreWeave GPU cloud. This workflow integrates unit testing, container image building (to GHCR), and rolling deployment to CoreWeave Kubernetes clusters via GitHub Actions. Ideal for ensuring stable, high-performance deployment of GPU-dependent models. Key features include validating GPU manifest requests and managing deployments using standard Kubernetes APIs.

A comprehensive guide to diagnose and resolve common CoreWeave infrastructure errors. It covers issues related to GPU scheduling, Pod status (e.g., Pending), CUDA Out of Memory errors, NCCL timeouts, and Kubernetes resource allocation within AI/ML deployment pipelines.

A comprehensive guide and set of tools for establishing a local development workflow tailored for CoreWeave GPU deployments. This includes steps for building Docker containers, validating Kubernetes YAML manifests using dry-run, and managing the full deployment lifecycle for AI inference services before pushing to the cloud.

A comprehensive utility designed for managing CoreWeave cloud infrastructure migrations. It facilitates upgrading between GPU generations (e.g., A100 to H100), updating CUDA versions, and handling Kubernetes resource changes. The tool uses K8s APIs to audit deployments, detect deprecated instances, and execute controlled rollbacks.

This workflow guides users through configuring CAST AI's autoscaler policies to achieve optimal cost management and resource utilization within a Kubernetes cluster. It covers enabling spot instances, setting node downscaler/evictor rules, defining cluster limits, and creating specialized node templates (e.g., for GPU or spot workers) using Terraform. Use this when implementing Phase 2 automation or optimizing cloud spend.

This bundle is designed to collect comprehensive diagnostics for CoreWeave clusters, including GPU node status, Kubernetes Pod logs, event history, and API connectivity. It aggregates critical cluster-level resource allocation and failure data into a single, actionable archive. Use it when preparing a support ticket to expedite the diagnosis of complex infrastructure issues like stuck pods, OOM errors, or resource bottlenecks.

This skill outlines how to implement and manage Role-Based Access Control (RBAC) and namespace isolation for multi-team GPU workloads running on CoreWeave's Kubernetes platform. It provides guidance on setting strict resource quotas, preventing noisy-neighbor issues, and defining granular roles (e.g., ML Engineer, Inference Operator). This is essential for maintaining resource isolation, ensuring compliance (SOC 2, HIPAA), and managing access across large-scale cloud deployments.

This guide details setting up and managing CoreWeave GPU cloud infrastructure across Development, Staging, and Production environments. It ensures strict resource separation using dedicated Kubernetes namespaces for GPU quotas and scaling policies. Use this when deploying complex ML/AI workloads requiring robust environment isolation and controlled promotion workflows.

This skill provides comprehensive observability for CoreWeave's GPU-intensive workloads running on Kubernetes. It monitors critical operational metrics, including GPU utilization, memory usage, inference latency (p99), and job completion rates. Use it for setting up DCGM metrics dashboards, configuring proactive alerts for underutilization or Out-Of-Memory (OOM) conditions, and ensuring high reliability for AI/ML deployments.