3. Data, Science & AI

Plans evolution of Qwen2.5-7B to SO8T model with Nobel-level math reasoning, advanced Japanese capabilities, and 2024-2026 LLM integration.

Tool for designing system prompts in AI agent systems, supporting modular construction and complex multi-agent orchestration with built-in patterns and dynamic research.

Enables Nobel Prize and Fields Medal level mathematical reasoning in SO8T models using alpha gate sigmoid control and golden ratio convergence.

Analyzes Magento 2 catalog structure, retrieves product data, and assesses catalog health for optimization and diagnostics.

Optimizes Magento 2 search performance and database efficiency through indexing strategies, Elasticsearch configuration, and database index design.

Compares two videos for quality and compression differences, generating interactive HTML reports with PSNR, SSIM metrics and frame-by-frame visual analysis.

Provides validation hooks for AI coding agents to ensure safe, deterministic code changes through automated tests, linters, and audit logs.

Transforms vague user inputs into structured, AI-ready prompts using EARS methodology to enhance LLM output quality for code, documents, and products.

Dispatches multiple AI agents to concurrently resolve independent tasks or failures, optimizing efficiency for 3+ simultaneous issues.

Prevents permanent data loss in beads (bd) issue tracker databases by providing safe recovery procedures for sync errors and critical table management.

Initializes Gemini CLI integration for the PACK auditor workflow, enabling AI-driven auditing capabilities.

Enables building AI applications with agent orchestration, structured outputs, and multi-provider LLM integration using the Dedalus SDK.

Diagnoses and optimizes slow database queries using explain plans, statistics, and index tuning to resolve performance bottlenecks.

Audits database data quality for nulls, duplicates, and invalid ranges, generating a report with remediation queries for debugging and validation.

Analyzes website traffic, engagement, and UTM campaign performance using Pirsch to identify visitor trends and top-performing pages.

Inspects database schema and metadata to aid in exploring, querying, and documenting database structures efficiently.

Manages context windows for LLMs to prevent bloat and maintain model performance efficiency.

Visually annotates and modifies matplotlib plots using color-based references for enhanced data visualization in scientific analysis.

Lightweight AI architecture playbook for dual-engine memory management (.ai_context) and manifest dispatching, optimized for Claude Skills bootstrapping and porting.

Lints PostgreSQL functions against schema, analyzes usage patterns, and generates repair reports for broken or unused database functions.

Specializes in synchronizing data between Meteor MongoDB and Supabase, including debugging sync failures and resolving entity mapping issues.

Provides secure, credential-managed access to production/staging databases for debugging, data verification, and RLS checks using environment variables and 1Password.

Manages Google Tag Manager configurations for analytics, conversion tracking, and event management on websites.

Queries production MongoDB databases for Meteor applications using environment credentials, enabling data verification and debugging of sync issues.