Data Science & ML Claude 技能

Senior Prompt Engineer

Optimizes LLM performance through advanced prompt engineering, RAG architecture design, and agentic system orchestration.

LLM CLI

Interacts with diverse large language models through a command-line interface to perform tasks like prompt execution, data extraction, and embedding management.

Llama Stack Code Reviewer

Streamlines code reviews for the Llama Stack repository by focusing on distributed system patterns, API compatibility, and automated testing fixtures.

Multi-Agent Architecture Patterns

Designs and implements sophisticated multi-agent architectures to overcome context limitations and handle complex task decomposition.

MOVA AI Bootstrap Generator

Generates static bootstrap packages to initialize MOVA AI models and environments without requiring external LLM calls.

Embedding Strategies for RAG

Optimizes embedding models and chunking strategies to enhance semantic search and RAG application performance.

RAG Implementation for LLM Apps

Builds high-performance Retrieval-Augmented Generation systems using vector databases and semantic search to ground AI responses in external knowledge.

Prompt Engineering Patterns

Implement advanced LLM prompting techniques like few-shot learning and chain-of-thought to enhance production AI reliability and output quality.

Book Style-Transfer SFT Pipeline

Converts literary works into high-quality supervised fine-tuning (SFT) datasets to train AI models in specific authorial voices.

ML Pipeline Workflow

Builds and orchestrates end-to-end MLOps pipelines from data preparation and model training to production deployment.

AI Project Development

Architects and manages the end-to-end lifecycle of LLM-powered applications using agentic development methodologies.

Market Analyst Brief

Generates concise, 24-hour tactical financial market briefs across multiple asset classes using real-time global news.

Context Optimization & Capacity Engineering

Optimizes AI context windows through strategic compression, masking, and partitioning to handle larger tasks and reduce operational costs.

LLM Performance Evaluation

Implement comprehensive evaluation frameworks for LLM applications using automated metrics, human feedback, and statistical benchmarking.

Marimo Reactive Notebook Development

Builds reactive Python notebooks and interactive data applications using marimo's directed acyclic graph (DAG) execution model.

Vector Search for Zettelkasten

Implements semantic search and AI-driven discovery for Obsidian vaults using vector embeddings and advanced chunking strategies.

Advanced Prompt Engineering

Optimizes and orchestrates advanced prompt engineering workflows for Claude 4.5 using industry-best patterns, guardrails, and context management.

Training Resilience for ML Trading

Optimizes Reinforcement Learning training by preventing premature stops and implementing adaptive recovery for trading models.

Trading Data Source Priority

Enforces high-quality Alpaca API data usage for crypto trading to prevent model failures caused by unreliable data fallbacks.

Bioinformatic Experiment Reporter

Synthesizes scientific lab notebooks into structured, publication-quality reports with AI-guided refinement and PDF export.

Deterministic Math Tools

Performs exact symbolic mathematical computations using SymPy to ensure accuracy without LLM estimation errors.

PyMC 5 Fundamentals

Facilitates the creation, review, and optimization of Bayesian statistical models using the PyMC 5 framework.

Bayesian Regression Models

Implements diverse Bayesian regression techniques using Stan and JAGS for advanced statistical modeling and uncertainty quantification.

RAG System Implementation

Build production-grade Retrieval-Augmented Generation (RAG) systems to ground LLM applications in external knowledge.

Databricks Genie Integration

Integrates Databricks Genie rooms into AI agent workflows to enable conversational BI and natural language data querying.

MLOps Pipeline Workflow

Builds production-grade MLOps pipelines by orchestrating data preparation, model training, validation, and automated deployment workflows.

RAG Implementation Guide

Implements robust Retrieval-Augmented Generation systems to connect LLMs with external knowledge bases and vector databases.

LLM Integration

Configures and manages diverse LLM providers, streaming callbacks, and token optimization strategies for XSky-based agentic workflows.

Memory Leak Detector

Identifies and resolves memory exhaustion issues in Python and PyTorch applications, specifically targeting matplotlib figures and tensor accumulation.

Architecture Reviewer

Enforces adherence to PRISM architectural patterns by validating the implementation of Runner and Trainer classes during refactoring.