3. Data, Science & AI

Advanced database design and administration for PostgreSQL, MongoDB, and Redis, including schema optimization, query tuning, and performance management.

Enables statistical analysis, machine learning model development, and business intelligence dashboard creation for data-driven insights.

Develops machine learning models and LLM integrations using PyTorch and TensorFlow for training pipelines and AI feature implementation.

Enables creation, editing, and analysis of spreadsheets with formula handling, formatting, and data visualization for multiple spreadsheet formats.

Generates executable code for precise numerical calculations and data transformations to eliminate errors in complex arithmetic and data processing tasks.

Verifies high-stakes decisions by generating multiple independent reasoning paths and confirming consensus to ensure accuracy before critical implementation.

Selects and applies the optimal AI reasoning framework for any task, optimizing performance and accuracy in AI-driven solutions.

Generates an HTML dashboard from meeting analysis to visualize insights and create summary reports.

Converts VTT transcript files into an interactive HTML dashboard with multi-tab analysis for meeting recordings and transcripts.

Manages database schema modifications, runs migrations, and deploys Supabase changes for SQL-based database operations.

Validates research, plans, and code for overengineering, underengineering, and AI hallucinations to ensure accuracy and quality.

Routes AI tasks to the most suitable LLM provider (Gemini, Codex, Copilot, Claude) based on task requirements.

Enables collaborative AI sessions with Codex (GPT-5.2) through Bash CLI, allowing named session management for AI task execution.

Builds stateful AI agents and agentic workflows with LangGraph in Python, supporting tool loops, branching, memory, and human oversight for production applications.

Builds AI agents using the Claude Agent SDK, supporting tools, hooks, subagents, MCP integration, and configuration in TypeScript/Python.

Provides PowerShell-based querying and analysis of Legislature (M2) QA database for legislative data including committees, bills, and meetings.

Converts data insights into executive-ready narratives using Gartner's What/Why/Next framework for strategic decision-making.

Enables real-time streaming of LLM responses for chat interfaces and applications across multiple AI providers with async iteration and usage tracking.

Provides a unified interface for building LLM applications with consistent message handling across OpenAI, Anthropic, Google, and Ollama, including lockfile management.

Enables querying and analysis of Eva QA database content including sessions, reports, and meeting notices using PowerShell.

Creates and manages the mandatory `llmring.lock` file for LLM applications, enabling model aliases, environment profiles, and fallback models.

Customizes retail AI agent configurations for diverse business types, industries, and analysis focus via prompt modification and vertical addition.

Searches articles using category filters and BM25-based keyword matching for precise information retrieval.

Allows the model to retain session context for more relevant and personalized responses.