3. Data, Science & AI

Integrates Google Analytics (GA4), Facebook Pixel, Google Tag Manager, and affiliate tracking for data collection and conversion monitoring.

Enables Claude-type AI to execute plans with zero errors, ensuring high-precision implementation of tasks.

Transforms text or documents into AI-generated infographics using Gemini Pro 3 for creation and Claude Vision for quality evaluation.

Reviews database migrations and schema changes, validating ID mappings, checking for data anomalies, and ensuring rollback safety for data integrity.

Provides deep analysis of codebases and market conditions using Google AI (Gemini) for insights without code modification.

Generates visualizations for machine learning experiment tracking using Weights & Biases (wandb) data.

Clusters metaphor annotations into four Conceptual Metaphor Theory (CMT) systems using quantitative analysis for linguistic research.

Builds Retrieval-Augmented Generation (RAG) systems by integrating vector databases to enhance AI model responses with external data.

Initializes a memory bank structure for AI context management in a new project workspace.

Manages plans and context documents through automatic indexing for efficient information retrieval and organization.

A unified AI agent for deep analysis, precise execution, and atomic-level investigation within the Polyprophet platform.

Orchestrates parallel chapter annotation for metaphor detection using subagents in the MIP pipeline.

Extracts text and structured data from PDF documents for data processing and analysis.

Optimizes LLM API costs via prompt caching, compression, and cost analysis without compromising output quality for AI applications.

Enables autonomous task execution and multi-agent orchestration using Google's Gemini AI model via command-line interface.

Detects market regimes (e.g., trending, volatile) and applies strategy-specific trading approaches based on real-time market analysis.

Reduces hallucinations and constraint violations in AI model planning through a two-phase reasoning approach for complex multi-step tasks.

Configures repository symlinks for AI tools (Gemini, Claude) to enable multi-agent compatibility and agent configuration via AGENTS.md linkage.

Enhances AI prompt effectiveness by applying CEV (coherent extrapolated volition) to align with user intent and goals.

Executes SQL queries on a local PostgreSQL database, supporting raw SQL and natural language input for data inspection and management.

Generates interactive HTML dashboards to visualize and analyze LLM usage statistics across multiple AI models and platforms.

Generates complex graph visualizations from Graphviz DOT code, providing both source files and pre-rendered images for technical data representation.

Automates web search results storage in Qdrant to build persistent knowledge bases for AI/ML applications like RAG.

CLI tool for interacting with Grok AI, enabling prompt submission and response retrieval to integrate with development tools like Claude Code.