Data Science & ML Agent Skills

Vercel AI SDK Documentation

Provides comprehensive documentation and implementation patterns for building AI-powered applications using the Vercel AI SDK.

Metabolomics Workbench Explorer

Accesses the NIH Metabolomics Workbench REST API to retrieve comprehensive metabolomics study data, metabolite structures, and standardized RefMet nomenclature.

MLOps Pipeline Workflow

Orchestrates end-to-end machine learning pipelines from automated data preparation through model training, validation, and production deployment.

LangGraph v1 & LangChain Assistant

Facilitates the development and migration of Python-based AI agents using LangGraph v1 and LangChain v1 standards.

Gemini CLI Integration

Integrates Google’s Gemini AI capabilities into Claude Code for multimodal tasks, advanced code analysis, and automated tool execution.

AC Memory Manager

Persists knowledge and learned patterns across autonomous coding sessions using Graphiti-based memory management.

LLM Tuning Patterns

Configures optimal LLM parameters and reasoning strategies based on evidence-backed research for specialized technical tasks.

Pymoo Optimization Framework

Optimizes complex multi-objective and constrained problems using state-of-the-art evolutionary algorithms and decision-making tools.

AC Memory Manager

Manages persistent cross-session memory for autonomous coding tasks using Graphiti integration.

Bedrock AgentCore Memory

Integrates persistent episodic and short-term memory into Amazon Bedrock agents to maintain user context and learn from interactions across sessions.

PufferLib Reinforcement Learning

Accelerates reinforcement learning workflows with high-performance environment vectorization and optimized PPO training.

NCBI Gene Database Integration

Simplifies gene data retrieval and annotation by querying the NCBI Gene database via E-utilities and Datasets APIs.

Scanpy Single-Cell Analysis

Performs comprehensive single-cell RNA-seq analysis workflows including quality control, clustering, and visualization.

scvi-tools: Probabilistic Single-Cell Analysis

Implement deep generative models for single-cell omics analysis, enabling batch correction, multimodal integration, and probabilistic differential expression.

LabArchives ELN Integration

Automates electronic lab notebook workflows including data uploads, notebook backups, and programmatic entry management via the LabArchives REST API.

FinnHub Financial Data

Integrates real-time financial market data including stock quotes, forex rates, crypto prices, and company fundamentals via the FinnHub API.

Python Experiment Reproducibility

Ensures deterministic results in Python research code by centralizing configurations, controlling randomness, and recording run metadata.

Agent Memory System

Implement a multi-layered persistent memory architecture for AI agents to enable cross-session learning and long-term context retention.

Gemini 3 Advanced Features

Implements advanced Gemini 3 Pro capabilities including function calling, structured JSON outputs, and cost-efficient context caching for production AI applications.

HMDB Metabolome Database

Accesses and queries the Human Metabolome Database for detailed metabolite information, clinical biomarkers, and biochemical pathways.

xAI Stock Sentiment Analysis

Analyzes real-time stock sentiment on X (Twitter) using Grok to provide actionable market insights and retail investor mood.

MATLAB & Octave Scientific Computing

Provides specialized support for MATLAB and GNU Octave numerical computing, matrix operations, and scientific visualization.

Gemini CLI for Claude Code

Integrates Google Gemini's multimodal reasoning and automated CLI tools into Claude Code for advanced AI collaboration and research.

COBRApy Metabolic Modeling

Performs constraint-based reconstruction and analysis (COBRA) of metabolic models for systems biology and metabolic engineering.

HS Risk Scoring & Reporting

Performs automated risk assessment on logistics data by identifying formatting errors, hazardous materials, and dimensional flags.

Bedrock Model Customization

Adapts Amazon Bedrock foundation models through fine-tuning, continued pre-training, reinforcement learning, and model distillation.

Bioinformatics & Genomic Analysis (pysam)

Processes genomic data files including SAM, BAM, VCF, and FASTA using a high-performance Python interface to htslib.

DrugBank Database Integration

Accesses and analyzes comprehensive pharmaceutical data, including drug properties, interactions, molecular targets, and chemical structures from DrugBank.

Gemini 3 Pro Advanced Features

Implements advanced Gemini 3 Pro capabilities including function calling, structured outputs, and context caching for production-grade AI applications.

Stable Baselines3 RL Suite

Implements production-ready reinforcement learning algorithms using a unified, scikit-learn-style API for efficient agent training and evaluation.