3. Data, Science & AI

Syncs Claude configuration to Codex and Gemini AI platforms for unified model settings.

Audits quality gates for large language models and identifies gaps to enhance performance and reliability.

Extracts structured content from web pages using the Firecrawl MCP web scraping service.

AI-powered web search and research tool using Perplexity's language model for contextual reasoning and information synthesis.

Builds and manages RAG pipelines for LLM applications, handling embeddings, retrieval, and flow orchestration.

Provides Scrapy-based web scraping, data parsing, and source management for structured data extraction and processing.

Provides analytics code navigation, metrics calculation, and dashboard data handling for portfolio performance tracking and error debugging.

Converts PDF, Word, Excel, images, audio, and other formats to Markdown for text analysis and LLM processing.

Batch-generates images via OpenAI Images API with random prompts and creates an HTML gallery for display.

Packages whole-codebase repositories for AI analysis within token limits, using folder-level reduction and Gemini feedback iteration.

Generates comprehensive ML/AI tutorial content emphasizing framework selection and implementation workflows for technical audiences.

Manages agent's conversational memory by saving, recalling, and organizing user-provided information for context-aware interactions.

Enables previewing and metadata inspection of Denmark statistics fact tables for data exploration and analysis.

Orchestrates multiple AI agents in parallel using Fan-Out/Fan-In pattern for efficient task distribution and result aggregation.

Explores Denmark's statistical subject hierarchy and fact table organization for data analysis and reporting.

Analyzes sales, customer, and inventory data to generate performance insights and forecasts for commerce operations.

Provides constraint satisfaction patterns tailored for the Pips puzzle solver, leveraging AI techniques to resolve complex puzzle configurations efficiently.

Guides integration of AI APIs such as OpenAI and ElevenLabs for application development.

Debugs vision API issues (Gemini/GPT) for puzzle image extraction, ensuring accurate analysis.

Validates AI prompt files against best practices to ensure quality in agent and skill configurations.

Specialist in deploying, fine-tuning, and monitoring LLMs with expertise in RAG pipelines, vector databases, and AI infrastructure maintenance.

Optimizes AEGIS model evaluation for ARC Challenge and GSM8K integrity checks through timeout analysis, robust answer extraction, and contamination detection.

Specializes in crafting and optimizing prompts for LLMs using advanced techniques like CoT, ReAct, and Few-Shot to maximize model reliability and performance.

Executes A/B/C tests compliant with official leaderboards, benchmarking AI models like Phi-3.5-mini-instruct with statistical validation.