3. Data, Science & AI

Creates and optimizes database schemas with Drizzle ORM, ensuring proper relationships, constraints, and indexes for efficient data management.

Converts GitHub, GitLab, and Gitee repositories into OpenCode Skills using embedded LLM processing with robust API handling for multiple mirrors and rate limits.

Provides GPU-optimized OCR for text extraction from images, PDFs, and multi-language documents with layout analysis and table detection.

Detects outliers and unusual patterns in business data including revenue, transactions, and customer behavior using statistical and machine learning techniques.

Designs end-to-end agentic systems including orchestration, RAG, tool contracts, and prompt strategy based on architecture artifacts.

Creates Entity-Relationship diagrams from database schemas to document and visualize database structure for technical teams.

Converts text to speech using ElevenLabs TTS for agent notifications and progress updates.

Configures MongoDB document storage, query operations, and indexing for Sorcha's distributed ledger, handling Register/Transaction/Docket data.

Configures AI/LLM model providers including OpenAI, Anthropic, Azure, and Bedrock for seamless integration in Spice applications.

Guides users on maintaining search index freshness, addressing stale data, outdated embeddings, and ensuring accurate search results in AI-driven systems.

Configures automatic task delegation between AI agents in Synapse A2A, supporting orchestrator and passthrough modes with status verification and error handling.

Manages agent memory for saving, recalling, and organizing user interactions to enhance conversational context and task continuity.

Creates and manipulates Excel spreadsheets with formulas, formatting, charts, and data analysis for efficient data handling and visualization.

Builds Model Context Protocol (MCP) servers to integrate external APIs and services with AI models, enabling context-aware interactions and tool augmentation.

Generates, parses, and analyzes StoBe DFT input files for quantum chemistry, XANES, and core-hole calculations.

Optimizes and designs effective prompts for LLM interactions, enhancing output quality and reliability in agent-based systems.

Guides creation and refinement of output-styles to customize Claude's behavior, personality, and persona design using best practices.

Simulates structured decision-making through AI advocate agents presenting opposing arguments for user evaluation and verdict.

A plugin for OpenCode that uses a persistent knowledge graph to enable AI agents to share findings, tasks, and context collaboratively.

Guides autonomous agent creation, structure, and system prompts for Claude Code plugins, including development best practices.

Provides techniques for evaluating and enhancing AI agent outputs via self-critique, rubric-based assessment, and iterative refinement pipelines.

Generates explicit data specifications from technical requirements, defining schemas, lifecycle, retention policies, and access controls for data contracts.

Provides PostgreSQL SQL best practices for writing clean, performant, and maintainable database queries and schema changes.

Structures multidisciplinary design reviews for LLM products, identifying actionable risks and gaps across UX, architecture, AI safety, and operations.