3. Data, Science & AI

Generates unified stock valuation reports using DCF, Reverse, Stress, Quality, and Masters methods with integrated data from Damodaran, Global Scouter, and yfinance.

Enforces dbt's SQL ownership principle for data transformations, ensuring SQL is handled in SQL, not Python.

Provides tools for building Cube semantic layers, including data modeling, dbt integration, and API validation for metrics and dimensions.

Enables building data lakehouse architectures using DuckDB, Iceberg, and Polaris catalog with dbt, Dagster, and REST catalog integrations.

Translates subtitles to target languages using Google's Gemini API with context-aware processing for accurate and natural translations.

Loads foundational principles for AI agents, providing a core set of axioms to guide agent behavior and decision-making.

Configures AI rule systems to set confidence thresholds and prevent false positives in CLAUDE.md and rule files.

AI-driven analysis of raw inputs to extract domain entities, workflows, and rules while filling knowledge gaps using market expertise in ontology building.

Orchestrates data pipelines via Dagster, enabling asset management, dbt integration, and compute abstractions across engines like DuckDB and Spark.

Configures Polaris catalog for data management, integrates with DuckDB and dbt, and resolves access control issues via OAuth2 and REST API research steps.

Verifies ontology pipeline outputs by cross-validating against inputs, checking traceability, and validating constraints to ensure data integrity.

Ingests and catalogs input materials from _input/ folder for initial ontology building phase, enabling domain analysis.

Provides essential utilities for managing Iceberg table storage, ACID transactions, time-travel queries, and schema evolution using the PyIceberg SDK.

Configures AI agent roles, handoff protocols, and model selection for agent-based systems.

Generates final ontology artifacts including entity definitions, workflow catalog, and concept guides from a Data Relationship Diagram (DRD) in the ontology building pipeline.

Designs and implements statistically valid A/B tests for product feature evaluation, ensuring robust experimental setup and significance analysis.

Enables implementation of analytics tracking systems to collect and analyze user interaction data for marketing insights and performance metrics.

Consults external AI models using Prompt150 formula (Context, Query, Method, Style) when internal sources are exhausted or confidence is low.

Command-line tool for web scraping and data extraction using Puppeteer and Playwright with rate limiting and ethical scraping practices.

Enables privacy-focused event tracking, analytics setup, and dashboard creation using common analytics tools.

Systematically audits document extraction pipeline quality through stratified sampling, verification, and reporting to ensure accuracy and reliability.

Creates mathematically precise time-series visualizations with technical validation, ensuring correctness of overlays, artifacts, and domain alignment.

Builds AI agents using the Claude Agent SDK in TypeScript, supporting batch workflows (V1) and interactive sessions (V2) with tool configuration and multi-turn patterns.

Retrieves health and fitness metrics from Apple Health, including activity, workouts, heart rate, and sleep data for user queries.