3. Data, Science & AI

Designs optimized database schemas for relational and NoSQL databases, including normalization, indexing, and performance strategies for PostgreSQL, MySQL, and MongoDB.

Specialized NLP processing for Chuukese language, including tokenization, accent handling, and cultural context preservation for translation and language model development.

Intelligently splits large texts into context-preserving chunks for AI training and RAG systems, maintaining semantic coherence and document structure.

Builds multi-agent systems with LangGraph, implementing supervisor, swarm, and hierarchical agent workflows for collaborative AI applications.

Guides users through initial data science project planning using Socratic questioning to clarify goals, data sources, and constraints.

Profiles data and generates structured analysis task breakdowns for subsequent data science phases in the DS workflow.

Provides structured checklists, confidence scoring, and issue identification for validating data analysis methodology, data quality, and statistical validity in Phase 4 of the /ds workflow.

Digitizes Chuukese text from scanned documents using AI-powered OCR with accent character and formatting preservation for accurate text extraction.

Processes multilingual documents with emphasis on low-resource languages, accented characters, and cross-linguistic analysis for minority and mixed-language content.

Generates high-quality training datasets from documents and text for AI models, optimized for low-resource languages and domain-specific knowledge.

Enables subagent delegation for data analysis tasks, dispatching specialized agents with output-first verification to ensure accuracy and efficiency.

Analyzes media files to extract and describe content including tables, text, and visual data using AI-powered processing.

Enables querying financial databases including WRDS, Compustat, CRSP, and SEC EDGAR filings via SQL patterns.

Manages Chuukese language linguistic data including dictionaries, phrase collections, translation pairs, and metadata through database operations.

Provides access to LSEG (Refinitiv) financial market data, including fundamentals, ESG scores, and RIC to ISIN conversion via the LSEG Data Library API.

Enables scalable document processing and LLM task execution via Gemini Batch API with production best practices.

Lists available data science plugins, tools, and data access skills (e.g., WRDS, LSEG) for user discovery and feature enablement.

Syncs Jupyter notebooks (.ipynb) with text-based files (.py) for version control, multi-kernel collaboration, and seamless data science workflow management.

Enforces output verification at every step in the Data Science workflow to ensure quality and correctness before advancing to subsequent phases.

Processes large documents (200+ pages) while preserving structure and extracting structured data from PDFs, DOCX, and text files with memory efficiency.

Enables database operations including querying, table management, and data processing in Apache Hive's data warehousing environment.

Step-by-step instructions for creating scheduled data processing jobs using the Hive framework.

Connects Spice to databases, cloud storage, and data warehouses for seamless data integration and source configuration.

Guides users through defining database schemas in Apache Hive, including table structures and data types for big data processing.