3. Data, Science & AI

Provides AI-powered web search, content extraction, and site mapping via Tavily's Python CLI for data gathering and analysis.

Stores user memories in a structured tripartite system (episodes, entities, skills) for AI memory management and retrieval.

Automates writing and updating BigQuery SQL queries and Python ETL scripts per Mozilla conventions, including test and schema coordination.

Manages BigQuery table metadata configurations including DAGs, schemas, and tests, coordinating with data pipeline tools to update configurations and tests when queries change.

Enables non-interactive batch processing of AI model calls for Gemini, Claude, and Codex, supporting translation, code review, and multi-model orchestration.

Coordinates multiple AI agents in parallel or sequential workflows, enabling task decomposition and delegation.

Processes audio interviews and PDF documents into structured data for data preparation stages.

Intercepts bulk file reads to replace them with semantic search, preventing token waste and optimizing AI processing efficiency.

Orchestrates AI model selection for coding tasks, balancing cost, batch processing, and API configuration to analyze multiple documents efficiently.

CLI tool for managing AI agent sessions, enabling session creation, status checks, and output retrieval.

A protocol for structured tracking of multiple external AI models during parallel execution, ensuring validation and complete results presentation.

Detects AI-generated text patterns using GLTR token analysis, DetectGPT curvature probes, and Coh-Metrix cohesion metrics for comprehensive writing assessment.

Triggers and manages memory consolidation within the World Weaver AI system to optimize memory usage and retrieval performance.

Orchestrates parallel AI agents for code analysis, sharing outputs to enable consensus and role-based task execution.

Validates code and analysis through parallel execution of multiple AI models (e.g., Grok, Gemini) with enforced performance statistics and consensus tracking.

Debugs tablature (TEF) files for banjo, resolving parsing errors, missing notes, and tuning issues from sources like Banjo Hangout.

Enables multi-strategy retrieval of stored memories from World Weaver for context-aware data access and analysis.

Automates pre-submission review for academic papers, including methodology checks, bias detection, AI text analysis, and statistical validation to ensure journal compliance.

Constructs comprehensive memory context from World Weaver for the current task.

Visualizes memory structures and generates diagrams for World Weaver systems to enhance understanding of complex memory systems.

Provides multi-engine speech-to-text transcription with real-time processing, speaker diarization, and multi-language support for audio files and meetings.

Manages a 5-phase AIDA pipeline through multi-agent delegation with Leader and Player subagents.

Executes trend analysis on recent project metrics including fix rate, NOT_PLANNED issues, and block patterns over a 3-day timeframe.

Step-by-step checklist and best practices for implementing AI agent tools, API routes, and capability extensions in the omer-akben portfolio.