Data Science & ML Claudeスキル

LLM Output Evaluation & Quality Assurance

Implements advanced LLM-as-judge patterns and RAGAS metrics to evaluate AI output quality and detect hallucinations.

Agent Loops & Reasoning

Implements autonomous agentic workflows and reasoning patterns for complex, multi-step LLM tasks.

LangGraph Agent Supervisor

Orchestrates multi-agent workflows using a central supervisor pattern to intelligently route tasks between specialized worker agents.

RAG Retrieval Patterns

Implements production-ready Retrieval-Augmented Generation patterns to ground AI responses in factual data and minimize hallucinations.

Agentic RAG Patterns

Builds self-correcting RAG systems using LangGraph for adaptive retrieval, document grading, and web search fallbacks.

Audio Language Models

Implements real-time voice agents, high-accuracy transcription, and expressive text-to-speech using native speech-to-speech models.

Prompt Engineering Suite

Optimizes AI application performance through production-ready prompt engineering patterns, versioning, and automated tuning.

LangGraph Conditional Routing

Implements dynamic workflow branching and retry logic for AI agentic systems using LangGraph patterns.

LangGraph Supervisor Pattern

Coordinates complex multi-agent workflows using a centralized supervisor-worker orchestration pattern.

Agentic RAG Patterns

Implements advanced Self-RAG and Corrective-RAG architectures for self-correcting AI retrieval systems.

Ollama Local Inference

Enables high-performance local LLM execution using Ollama to eliminate API costs and enhance data privacy during development.

Agentic Workflows & Loops

Implements autonomous reasoning patterns like ReAct and Plan-and-Execute to enable LLMs to solve complex, multi-step tasks.

Semantic Redis Caching

Optimizes LLM application performance and costs using vector-based similarity caching with Redis.

Prompt Caching Optimizer

Optimizes LLM API costs and performance by implementing provider-native prompt caching for Claude and OpenAI.

Mem0 Long-Term Memory

Persists semantic context and project decisions across multiple Claude Code sessions using Mem0.

LangGraph State Management

Architects and manages sophisticated state schemas for LangGraph workflows using industry-standard patterns and performance optimizations.

Vision Language Models

Integrates advanced multimodal capabilities from leading AI models for image analysis, document understanding, and visual reasoning.

Context Engineering

Optimizes Claude Code's context window through attention-aware positioning, progressive disclosure, and strategic token budgeting.

Reranking Patterns for RAG

Improves search precision and retrieval quality in RAG pipelines by re-scoring documents with cross-encoders and LLMs.

LangGraph Parallel Execution

Implements high-performance parallel execution patterns in LangGraph to enable concurrent agent workflows, fan-out/fan-in structures, and efficient map-reduce operations.

Golden Dataset Curation

Curates and validates high-quality AI evaluation datasets using multi-agent analysis and automated quality scoring.

HyDE Semantic Retrieval

Improves semantic search accuracy by generating hypothetical answer documents to bridge vocabulary gaps in RAG pipelines.

Contextual Retrieval RAG

Enhances RAG pipelines by prepending situational context to document chunks to preserve semantic meaning and significantly improve retrieval accuracy.

Fine-Tuning & Customization

Optimizes Large Language Models for specific domains using parameter-efficient fine-tuning, DPO alignment, and synthetic data generation.

LangGraph Functional API

Builds production-ready AI workflows using Python decorators for task orchestration and parallel execution.

Ollama Local Inference

Powers local LLM integration and performance tuning for private, cost-effective AI development using Ollama.

LangGraph State Management

Designs and manages robust state schemas for AI agent workflows using LangGraph best practices and modern design patterns.

Memory Fabric

Orchestrates unified AI memory by combining local knowledge graphs with semantic cloud search for persistent, high-context retrieval.

Embeddings Integration

Implements high-performance text embeddings for semantic search, document similarity, and vector database integration.

Prompt Engineering Suite

Optimizes LLM performance through production-ready patterns including Chain-of-Thought, dynamic few-shot learning, and automated prompt tuning.