3. Data, Science & AI

Routes users to the appropriate AI agent by analyzing project context, optimizing multi-agent system interactions.

Provides structured prompt engineering standards and context engineering principles for AI agents, optimized for clarity and signal-to-noise ratio based on Anthropic best practices.

Provides context-efficient, read-only delegation to specialized subagents within AI systems for optimized task handling.

Designs scalable database schemas for SQL and NoSQL, including normalization, indexing, and performance optimization to ensure data integrity and efficiency.

Memory management system for AI agents that recalls context before tasks, records learnings after, and reviews recommendations for continuous improvement.

Accelerates AI-driven prototyping cycles using fast inference for tight feedback loops on POCs, spikes, and concept exploration.

Expert assistance in writing optimized SQL queries, interpreting execution plans, and designing efficient database schemas.

Optimizes LLM prompts using structured techniques, transforming vague requests into clear, effective instructions with examples and constraints.

Analyzes datasets to uncover patterns, recommends visualizations, and delivers data-driven insights for strategic decision-making.

Guides users in optimizing LLM prompts via DSPy code and custom GEPA algorithm implementation based on dataset and grader requirements.

Generates truncated evaluations to simulate different context window limits for LLM performance benchmarking.

Imports Hugging Face model evaluation results into ReasonScape m12x dataset for performance analysis.

Provides database management commands and Ruby script execution for querying and manipulating event databases without a Rails server.

Creates clean, data-focused charts and dashboards using compact JSON specifications for efficient visualization.

Enables integration with Tautulli for Plex streaming analytics, including data fetching, watch history analysis, and user activity pattern visualization.

Optimizes structured data using TOON format, reducing token count by 30-60% for arrays, tables, logs, and API responses.

Expert in retrieving and managing banking transaction data via Plaid Transactions API, including sync, categorization, and historical data.

Provides expert guidance on Anthropic Claude API usage, including model selection, prompt engineering, function calling, and vision capabilities for AI applications.

Retrieves OpenShift release dates and metadata via the Sippy API for tracking and analysis purposes.

Expert integration of Plaid banking API for financial data, including account authentication, transaction history, balance checks, and ACH transfers in fintech applications.

Analyzes BigQuery usage patterns, costs, and query performance to optimize data operations and reduce expenses.

Formats structured data (tables, logs, transactions) in TOON v2.0 with customizable arrays and delimiters for AI/ML pipelines and data processing.

Grades component health using regression triage metrics for OpenShift releases, enabling data-driven triage decisions.

Connects to AI agents via A2A protocol, enabling message exchange, response handling, and real-time result streaming for agent-based workflows.