Data Science & ML Claude 技能

Hardhat R Modeling Infrastructure

Standardizes the creation of R modeling packages by providing consistent preprocessing interfaces and output formatting patterns.

Tidymodels Overview for R

Streamlines machine learning workflows in R using the consistent and modular tidymodels ecosystem.

R Metaprogramming with rlang

Manipulates R expressions and builds dynamic code using rlang's defuse and inject mechanics.

Complex R Targets Pipelines

Orchestrates advanced R data pipelines using complex branching patterns, custom target factories, and efficient iteration strategies.

Tidy R Function Design

Designs and reviews R function APIs to ensure they are predictable, pipe-friendly, and follow Tidy Design Principles.

AI & Emerging Technologies

Builds sophisticated AI-powered applications and blockchain solutions using modern LLMs, machine learning frameworks, and Web3 technologies.

Google OR-Tools Optimizer

Solves complex combinatorial optimization problems including vehicle routing, scheduling, and resource allocation using Google's open-source suite.

Pivot Pipeline Stage Writer

Streamlines the creation of Pivot data pipeline stages using Python type annotations for automated file I/O and caching.

AWS FDE for AI/ML

Builds production-ready AI/ML prototypes and POCs on AWS with cost optimization and Well-Architected best practices.

ML Training Debugger

Diagnoses and resolves machine learning training failures like loss divergence and gradient issues through automated artifact analysis.

Dataset Catalog Review

Audits and validates dataset catalog entries to ensure metadata consistency, template compliance, and RAG-readiness.

Component Model Analysis

Evaluates framework extensibility, abstraction layers, and configuration patterns to assess architectural flexibility.

Data Wrangler

Transforms and analyzes large datasets using DuckDB SQL directly within the Claude Code environment.

DuckDB for Scientific Data

Integrates high-performance analytical SQL capabilities into Python workflows for efficient data processing and large-scale querying.

Data Engineering & AI Systems

Accelerates the development of data pipelines, machine learning models, and modern AI agent architectures.

Trading Strategy Research

Streamlines the development, validation, and systematic documentation of trading strategies and market edges.

Gemini Real Adapter

Facilitates direct integration with the Google Gemini API for multi-modal content handling and context management within multi-model infrastructures.

Multi-Provider Model Orchestrator

Orchestrates multiple AI model providers to optimize development workflows for cost, latency, and reasoning capability.

ML Fundamentals

Master core machine learning pillars including data preprocessing, feature engineering, and robust model evaluation pipelines.

Scientific Visualization

Generates publication-quality scientific figures and multi-panel plots adhering to journal standards and accessibility guidelines.

Executing HEC-HMS Runs

Automates HEC-HMS hydrologic simulations with support for parallel batch processing and multi-version compatibility.

Denmark Statistics Subject Explorer

Navigates and explores the Denmark Statistics (DST) subject hierarchy to discover specific fact tables and data categories.

Cook County Data Portal

Enables seamless querying and extraction of datasets from the Cook County Open Data Portal using SODA and SoQL.

Chicago Data Portal Explorer

Facilitates seamless querying and downloading of City of Chicago municipal datasets using the Socrata Open Data API and SoQL.

US Census Data Query

Streamlines the retrieval of demographic, economic, and housing data from the US Census Bureau API.

KINTSUGI SLURM Channel Naming

Standardizes SLURM job output naming by mapping channel numbers to biological marker names for the KINTSUGI pipeline.

GGUF Model Quantization & Inference

Optimizes and deploys large language models using GGUF format and llama.cpp for efficient inference on consumer hardware.

XGBoost & LightGBM ML Toolkit

Implements industry-standard gradient boosting algorithms for high-performance machine learning on tabular and structured datasets.

CSV Data Analyzer

Analyzes CSV datasets using Python pandas to extract statistical insights, detect patterns, and identify data quality issues.

CODEX Experiment JSON Parser

Translates CODEX/Akoya experiment.json metadata into the KINTSUGI ExperimentConfig format with precise field and wavelength mapping.