3. Data, Science & AI

Executes SQL queries across Snowflake environments (QA, staging, production) for data validation and access to WS_HHH_BI_DW.

Plans and implements safe database schema changes including migrations, indexes, and backfills for tables, columns, and ORM relationships.

Validates historical data completeness and quality across specified date ranges to ensure data integrity and reliability.

Adapts AI behavior to task types using predefined operational modes (brainstorm, implement, debug) for optimized task execution.

Fetches and organizes Pokemon data from PokeAPI by generation, providing structured JSON output for easy integration.

Analyzes images, audio, and video using Gemini API, generates images with Imagen 4 and videos with Veo 3 for multimodal AI applications.

Optimizes LLM token usage by managing session context and research cache, ensuring efficient resource allocation during interactions and research operations.

Spawns an independent agent process to handle complex sub-tasks, enabling recursive agent-based task decomposition within LLM workflows.

Validates AI agent configurations for model selection, tool permissions, and approach quality to ensure optimal performance and adherence to best practices.

Formal framework for evaluating LLM features using Evaluation-Driven Development (EDD) principles to ensure robust model performance.

Designs optimized prompts for LLM applications, covering structure, context, and output formatting to enhance model performance.

Assesses code repository readiness for AI agent integration by evaluating 81 criteria across 8 pillars, assigning maturity levels and generating visual reports.

Optimizes Claude Code's context window to prevent token limit issues and ensure smooth LLM interactions.

Provides expert machine learning engineering context and support for Lumina project development, including model training and deployment.

Enables creation, editing, and analysis of .xlsx files with formula preservation and data visualization.

Warms up LLM context with foundational knowledge to enhance subsequent AI interactions and response quality.

Optimizes Claude model selection (Opus/Sonnet/Haiku) based on task complexity and context for quality and cost efficiency.

Translates Splunk SPL queries to Axiom APL, providing command mappings and syntax transformations for data query migration and learning.

Guides auditing, refactoring, and production design of ML Jupyter notebooks, addressing anti-patterns, data leakage, and reproducibility.

Extracts structured data from IC/component datasheets (PDFs/URLs) and generates consistent markdown summaries for hardware design documentation.

Searches 1000+ curated AI/ML project summaries on GitHub with bilingual support, relevance scoring, and language filtering.

Analyzes assessment data to calculate mastery rates, identify achievement gaps, and generate learning analytics dashboards for educational outcomes measurement.

Enables business intelligence through SQL queries on data warehouses to answer metrics, trends, and data-driven questions.

Generates declarative multi-agent systems (MAS) using POMASA pattern language for AI agent orchestration and research automation workflows.