Skill UI

This skill provides comprehensive guidance on training Mixture of Experts (MoE) models, including architectures like Mixtral and DeepSeek-V3. It details techniques for achieving large-scale model capacity with significantly reduced computational costs (e.g., 5x reduction vs. dense models). Topics covered include top-k routing, load balancing, expert parallelism, and integration with frameworks like DeepSpeed and HuggingFace for efficient, resource-constrained training.

Delivers end-to-end monitoring setups—structured logging, Prometheus/Grafana dashboards, alerting rules, distributed tracing, and load/profile testing—so teams can observe services, debug production issues, and plan capacity.

This skill provides comprehensive guidance on load testing and capacity planning for the Perplexity Sonar API. Users learn how to simulate high traffic volumes, respecting strict rate limits (like 50 RPM), benchmark latency, and estimate operational costs. It is essential for performance engineering, capacity architecture design, or stress testing APIs at scale.

This comprehensive guide provides detailed steps for load testing and capacity planning of Replit deployments. Learn how to benchmark performance using industry tools like k6 and autocannon, optimize Autoscale behavior, right-size Reserved VMs, and manage database connection pools. It covers optimizing cold start times and ensuring high availability under variable and peak traffic loads.

This guide provides comprehensive methods for load testing and capacity planning of Vercel serverless deployments. Learn how to use professional tools like k6 and autocannon to simulate realistic traffic spikes, measure latency (P95, P99), analyze cold start impact, and tune concurrency settings (Fluid Compute) to ensure optimal application performance and scalability under heavy load.

SimPy is a powerful, process-based discrete-event simulation framework written in Python. It is designed for modeling complex systems where entities interact with shared resources and time-based events occur, such as manufacturing lines, service queues, network traffic, or logistics networks. Use this skill to analyze resource contention, simulate process optimization, conduct capacity planning, and model dynamic system behaviors.

This skill provides sophisticated project estimation by adapting to human-only, hybrid, or agent-first development modes. It uses research-backed PERT statistics and confidence bands (P50, P75, P90) to forecast development timelines for tasks, backlogs, and capacity planning. It continuously calibrates estimates based on actual completion data, ensuring accuracy in AI-assisted and hybrid development environments.

A behavioral psychologist skill designed to ethically engineer genuine scarcity and urgency mechanics. This tool helps structure compelling messaging that motivates immediate action by leveraging real limitations (inventory, time, capacity) rather than misleading claims. It ensures persuasive communication remains trustworthy and preserves audience autonomy.

This comprehensive guide dives deep into the internals of Go's built-in data structures (slices, maps, arrays) and standard library packages. It covers critical topics like capacity growth mechanics, preallocation strategies, pointer types (unsafe/weak), and the optimized usage of specialized containers (heap, list, ring). Use this resource to select the correct structure for optimal performance and memory safety in expert Go development.

A generic, type-safe in-memory caching library for Go (Go 1.22+). It offers advanced features like nine specialized eviction algorithms (LRU, LFU, W-TinyLFU), Time-To-Live (TTL), read-through loaders, and Prometheus metrics. This skill helps developers design robust caching layers, optimize performance by selecting the correct eviction policy based on access patterns, and manage cache capacity effectively.

Provides systematic Power BI performance troubleshooting guidance covering model optimization, DAX design, report layout, and capacity diagnostics, plus tool-driven measurement tips for faster page loads, interactions, and query execution.

Detects host inventory including CPU, memory, disk, and accelerator limits for resource-aware planning. Generates a redacted JSON snapshot without stress tests. Supports Python 3.11+ on Linux, macOS, and Windows. Includes workload planning helpers and conservative capacity estimation.