Data Science & ML Agent Skills

Prompt Engineer

Architects high-performance system instructions and prompt patterns to optimize Large Language Model outputs and consistency.

Genomic Machine Learning (Geniml)

Builds machine learning models and unsupervised embeddings for genomic interval data and BED file collections.

Academic Research Engineer

Implements high-performance systems and complex algorithms with absolute scientific rigor and formal correctness.

Project Designer

Transforms initial software concepts into comprehensive, well-architected technical designs and implementation roadmaps.

Voice AI Development Architect

Builds low-latency, production-grade voice applications using real-time APIs, advanced synthesis, and streaming transcription services.

PyTorch Docstring Writer

Generates and updates professional PyTorch-style docstrings using Sphinx and reStructuredText conventions.

Scientific Schematics

Generates publication-quality scientific diagrams and technical schematics with automated AI quality review and smart iteration.

Reference Indexer

Indexes and manages external reference documents to power Retrieval-Augmented Generation (RAG) for domain-specific AI workflows.

Math Intelligence Router

Orchestrates complex mathematical computations by deterministically routing natural language requests to specialized CLI tools for symbolic math, logic, and geometry.

Senior Data Engineer

Builds scalable, production-grade data pipelines and ETL/ELT systems using the modern data stack.

Adaptyv Protein Lab

Connects Claude to cloud laboratory services for automated protein testing, sequence optimization, and wet-lab validation.

PyTorch Geometric (PyG) Support

Enables development and training of Graph Neural Networks (GNNs) using the PyTorch Geometric library.

bioRxiv Research Database Search

Searches and retrieves life sciences preprints from the bioRxiv database with advanced filtering and PDF download capabilities.

Perplexity AI Search

Executes real-time, AI-powered web searches with cited sources and scientific literature access using Perplexity models.

GWAS Catalog Database

Queries the NHGRI-EBI GWAS Catalog to retrieve SNP-trait associations, genetic variant data, and genome-wide association study summary statistics.

Research Question Refiner

Transforms vague research interests into concrete, actionable, and tractable research questions with a systematic feasibility analysis.

HMDB Database Explorer

Accesses the Human Metabolome Database (HMDB) to retrieve metabolite data, chemical properties, and clinical biomarkers for metabolomics research.

Experiment Design Checklist

Generates rigorous experimental frameworks for scientific research and machine learning projects to ensure statistically significant and defensible results.

Hugging Face Space Deployer

Deploys machine learning models to Hugging Face Spaces using optimized configurations for Gradio, ZeroGPU, and LoRA adapters.

Research Question Refiner

Transforms vague research interests into concrete, measurable, and tractable research questions through systematic refinement and feasibility analysis.

FDA Database Access

Simplifies querying the openFDA API to analyze regulatory data, drug safety profiles, medical device clearances, and food recalls.

Agentic Design Patterns

Provides architectural patterns and implementation guides for building reliable autonomous AI agent systems.

Strategy Optimization for Finance

Implements a systematic methodology for diagnosing, refining, and validating trading strategies to improve win rates and returns.

Research & Professional Analysis

Provides structured methodologies and frameworks for market research, competitor analysis, and professional data synthesis.

LangGraph Python Expert

Architects and implements sophisticated, stateful multi-agent LLM applications using LangGraph and Python.

NLP Basics

Simplifies text analysis and processing using modern NLP techniques including embeddings, tokenization, and transformer models.

Snakemake Per-Channel Resume Logic

Implements granular skip-existing checks in Snakemake wrapper scripts to resume interrupted HPC jobs without re-processing completed channels.

KINTSUGI GPU-Only SLURM Scheduling

Optimizes KINTSUGI batch processing by enforcing GPU-only SLURM scheduling to achieve up to 25x speedups over CPU fallback.

Weighted Autofluorescence Subtraction

Enhances multiplex immunofluorescence images by applying range-specific weights to remove background noise while preserving delicate biological signals.

Gemini CLI Integration

Integrates the Google Gemini CLI into Claude to provide large-context analysis, safe sandbox execution, and structured code modifications.