Data Science & ML Agent Skills

Advanced LLM Evaluation

Implements production-grade LLM-as-a-judge patterns to evaluate model outputs with high reliability and bias mitigation.

Excel & Data Analysis Toolkit

Creates, edits, and analyzes Excel spreadsheets with professional formatting, automated formula recalculation, and integrated data visualization.

Advanced LLM Evaluation

Implements sophisticated LLM-as-judge methodologies to evaluate and compare AI model outputs with high reliability and bias mitigation.

DSPy.rb Expert

Builds type-safe, composable LLM applications in Ruby using the DSPy framework to program AI behavior instead of manual prompting.

Google Gemini Embeddings

Builds production-ready RAG systems and semantic search using optimized Gemini embedding-001 models and vector storage patterns.

Polars Data Engineering

Accelerates data manipulation and ETL pipelines with the high-performance Polars DataFrame library.

SHAP Model Explainability

Explains machine learning model predictions and feature importance using SHAP values and comprehensive visualizations.

Stonk: Ethical Investment Intelligence

Analyzes global investments through the lens of power structures, ethical constraints, and geopolitical alignments.

AI Gateway Manager

Manages and routes requests across multiple AI providers including Anthropic, Ollama, and HuggingFace.

RAG Implementation

Implements Retrieval-Augmented Generation (RAG) systems with vector databases and semantic search to build grounded, knowledge-aware AI applications.

Gemini Image Generator

Generates high-quality images from text prompts using Google Gemini 3 Pro via the fal.ai API.

IGV Integration

Automates IGV snapshot generation for visualizing genomic alignments and variant calls in BAM files.

Flow Nexus Neural Networks

Trains and deploys complex neural network architectures within distributed E2B sandbox environments for scalable machine learning workflows.

COSMIC Cancer Gene Toolkit

Queries and annotates genomic data using the COSMIC Cancer Gene Census to identify known cancer genes and their clinical properties.

VCF Variant Analysis Toolkit

Analyzes, filters, and exports genomic variant data from VCF and BCF files for bioinformatics and sequencing workflows.

Scikit-learn Machine Learning

Builds, evaluates, and deploys production-ready machine learning models using the industry-standard scikit-learn library.

Structural Biology Analysis

Analyzes protein and molecular structures through AlphaFold interpretation, quality validation metrics, and comparative structural techniques.

BAM Alignment Toolkit

Analyzes genomic alignment files to extract reads, identify insertions and deletions, and calculate coverage statistics for WGS and WES data.

Local BLAT DNA Alignment

Performs high-speed local DNA sequence alignment against hg38 and CHM13 genomic references without external API dependencies.

NCBI BLAST Search

Performs NCBI BLAST sequence similarity searches using BioPython to identify homologous DNA or protein sequences.

Biological Sequence I/O

Reads, writes, and manipulates biological sequence data formats like FASTA, FASTQ, and GenBank.

AgentDB Learning Plugins

Implements and trains advanced reinforcement learning algorithms to create autonomous agents that evolve through experience.

HPC Development & Testing Workflow

Streamlines scientific development on HPC environments using multi-root workspaces and automated test data extraction.

Prompt Engineering Lab

Creates, optimizes, and debugs high-performing, production-ready prompts for Claude 4, GLM 4.7, and Gemini 3 using evidence-based techniques.

PyTorch Geometric (PyG)

Develops and trains Graph Neural Networks (GNNs) for node classification, link prediction, and geometric deep learning tasks.

Hybrid Search Implementation

Combines vector similarity and keyword-based search to optimize retrieval accuracy in RAG systems and search engines.

Similarity Search Patterns

Implements high-performance similarity search and vector database patterns for semantic retrieval and RAG systems.

Plotly Interactive Visualization

Generates interactive, publication-quality Python charts and dashboards for data exploration and presentation.

GPU Parallel Scheduling & Memory Safety

Implements queue-based GPU allocation and memory cleanup patterns to prevent OOM crashes and ensure reliable progress tracking in parallel workflows.

Prompt Engineering Patterns

Optimizes Large Language Model performance through advanced reasoning patterns, few-shot learning, and structured prompt templates.