3. Data, Science & AI

Systematically tests, evaluates, and compares LLM prompts to ensure quality and effectiveness in AI model interactions.

Optimizes complex SQL queries and patterns for PostgreSQL, MySQL, and SQL Server databases.

Coordinates multiple AI models to collaborate on complex technical tasks including debugging, code reviews, and architectural analysis.

Framework for safe A/B testing with statistical validation to optimize product features and marketing strategies through data-driven decisions.

Optimizes AI agent prompts via genetic algorithms and historical performance data to improve model effectiveness.

Forecasts potential issues using trend analysis and pattern matching to enable proactive resolution.

Analyzes telemetry data to identify and extract patterns in software development processes.

Analyzes historical patterns to generate data-driven improvement suggestions with ROI calculations for prioritized business decisions.

Executes spatial database operations for aviation data, including indexing, proximity queries, and geometric calculations on features like airports and airspace.

Georeferences FAA d-TPP charts and airport diagrams to enable location-aware overlays and correlate CIFP procedures with chart imagery for aviation navigation systems.

Automates parsing of FAA aviation data sources (NASR, CIFP, etc.) into SQLite databases for AIRAC cycle updates and data pipeline integration.

Builds Model Context Protocol (MCP) clients to connect AI agents with external tools and data sources via MCP servers.

Fetches financial data from LSEG Workspace using Eikon Data API, handling cURL requests, field discovery, and time series retrieval.

Fetches and processes oil/gas trade flow data including trade queries, aggregations, vessel positions, and company fleet information from Kpler.

Manages LLM API keys, supports multiple AI providers, and orchestrates AI agents for seamless integration.

Self-evolving AI agent now operating on atomic architecture for modular, scalable adaptation and continuous learning.

Generates images for worldbuilding entities using OpenAI's gpt-image-1.5 model, processing prompts from entity files and updating references.

Enables setup, improvement, and auditing of analytics tracking implementations including GA4, event tracking, and UTM parameters.

Provides 60+ schema-driven jq transformations for efficient JSON manipulation, filtering, and validation in data workflows.

Analyzes agent failures to refine agent definitions and improve performance within the current conversation session.

Analyzes content format suitability across Film, TV Series, Animation, Microdrama, and Audio Drama using GPT-4o for batch processing and catalog reporting.

Manages OpenAI embeddings for title vector search, enabling batch regeneration, single updates, verification, and cost estimation for improved search accuracy.

Automates batch collection of views and ratings data from Korean webtoon/webnovel platforms (Naver, Kakao, Manta) and ingests it into a titles database for metrics tracking.

Automates batch generation of Hollywood comparable titles via GPT-4o for catalog coverage and overnight processing.