Skill UI

This skill set provides advanced, structured strategies for optimizing Large Language Model (LLM) interactions. It covers techniques such as intelligent context summarization, dynamic trimming, context routing, and precise token budget allocation to prevent 'context rot'. It is essential for building robust, long-running, multi-turn conversational AI systems.

Combines model routing, immutable budget tracking, retry policies, and prompt caching so LLM API calls stay on-budget without sacrificing quality, ideal for batch workloads with mixed complexity.

Guide to reducing Groq inference spend by routing requests to cost-effective models, trimming token usage, caching repeated calls, batching requests, and setting spend limits for Groq Cloud billing scenarios.

Defines best-practice Groq deployment with tiered model routing, middleware, streaming pipelines, and fallback chains for ultra-fast LLM inference and production monitoring when launching new Groq integrations.

Provides code for tracking per-minute request and token limits with rate limiter, retry/backoff loops, model routing, and batching to keep Mistral AI usage within tiers and avoid 429 errors.

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.

Implements advanced OpenRouter model routing with A/B testing, gradual rollouts, and performance-aware selection to optimize multi-model deployments and experiments across environments.

Implements intelligent request-based routing logic to choose OpenRouter models that balance cost, speed, and quality, guiding teams through verification, environment adaptation, and monitoring.

Optimize Perplexity AI search billing by routing queries to the right model, caching responses, limiting token output, filtering domains, and monitoring usage so your API spend stays predictable.

Sets up Perplexity Enterprise RBAC with SSO, organization controls, and model-restricted API keys for teams, includes gateway routing, domain filtering, usage monitoring, and key rotation routines.

Sets up Perplexity Sonar API clients across development, staging, and production by routing sonar vs. sonar-pro models per workload, enforcing environment-specific limits, and controlling costs for multi-environment search deployments.

Defines a Perplexity reference architecture for AI research and search projects, covering model routing, citation extraction, conversational search, and multi-query pipelines to establish consistent integrations and reviews.