3. Data, Science & AI

Manages cross-agent communication with semantic search, deduplication, and GitHub Issues synchronization for seamless AI workflow continuity.

Optimizes large language model performance in production using advanced prompt engineering techniques for reliability and control.

Efficiently manages AI model context windows by summarizing, focusing, and pruning text to remain within token limits, optimizing LLM interactions.

Creates specialized Claude Code sub-agents with customizable system prompts and tool configurations for specific AI task workflows.

Provides core database utilities for the ikigai project, enabling data querying, management, and relational operations.

Compares Xparcel Expedited and Ground rate sheets year-over-year, generating Excel reports with side-by-side rates, differences, and summaries.

Detects recurring patterns in issues and events, identifying temporal, resource-based, cluster, and cascading patterns to suggest prevention strategies.

Analyzes unstructured shipping data, preprocesses and cleans it, then generates dashboards with volume, cost, and savings insights for carrier operations.

Analyzes shipping data exports from FreightClub and ShipStation to assess rates and identify FirstMile opportunities.

Analyzes shipping datasets to generate reports on DIM exposure, zone distribution, and cost intelligence using normalized address grouping and conditional analytics.

Executes AI agent evaluations via YAML test suites with automated, LLM-as-judge, and human review capabilities.

Analyzes shipping data exports to understand customer operations and inform rate creation, focusing on discovery before proposal.

Collects and filters AI-focused news articles from multiple RSS feeds, including tech blogs and research sources, based on age and relevance.

AI-driven continuous learning system using Critic and Reflection Agents with Discord-based approval workflow for skill proposals, Voyager-inspired.

Analyzes article collections to identify hot topics, track momentum, and detect emerging themes using AI-driven data science techniques.

Enables high-performance data virtualization for massive datasets while maintaining optimal user experience through efficient data access strategies.

Analyzes raw articles using LLM to generate summaries, extract entities, categorize content, assign importance, and detect breaking news.

Performs context-aware memory search across hierarchical tiers (working, episodic, semantic) to retrieve relevant information and match patterns.

Enables high-performance rendering of large datasets as interactive charts and graphs in React using Canvas, SVG, and WebGL.

Creates and configures specialized Claude Code sub-agents with optimal settings for task-specific AI workflows.

Executes PostgreSQL SQL queries for data analysis, schema inspection, and query optimization.

Provides a command-line interface to interact with Model Context Protocol (MCP) servers for accessing external tools, APIs, and data sources in AI applications.

Optimizes text prompts for Gemini image generation to ensure high-quality, model-compliant outputs before API calls.

Enables distributed neural network training, deployment, and lifecycle management in Flow Nexus cloud for scalable AI solutions.