3. Data, Science & AI

Designs scalable data systems using Pat Helland's patterns, emphasizing idempotency and distributed data beyond ACID transactions for high-scale applications.

Builds systematic trading systems and quantitative research platforms using rigorous data analysis and computational finance methodologies inspired by D.E. Shaw.

Designs fault-tolerant data systems using ACID properties and transaction processing principles for databases and payment systems.

Builds scalable ML pipelines for financial trading systems, emphasizing feature engineering, distributed computing, and rigorous infrastructure for alpha research and backtesting.

Builds quantitative investment systems using factor models, emphasizing academic rigor and systematic portfolio design for asset pricing research and factor-based investing.

Optimizes database performance using rigorous benchmarking and internal understanding, following Andy Pavlo's methodology for query tuning and storage engine design.

Provides statistical trading strategies based on mean reversion, utilizing standard deviation bands and RSI extremes for systematic market entry and risk management.

Enables iterative hypothesis validation and feedback-driven model refinement for data-centric AI applications and experimental workflows.

Queries and displays Claude Code usage statistics from the vibe-check database, including total usage and session details.

Slots external knowledge into Rune constructs for enhanced AI model responses and knowledge augmentation.

Analyzes and presents frequency of tool usage in user-Claude interactions, delivering usage statistics for workflow optimization.

Generates structured physics reference tables and explains timing rationale for scientific calculations and engineering applications.

Executes physics-based and rigorous validation checks on sprint components.

Assesses codebases for autonomous agent development readiness, providing tailored recommendations on infrastructure, code practices, and agent workflows.

Architects and debugs hierarchical matching systems using Gale-Shapley and Hungarian algorithms for two-sided matching, optimization, and entity resolution.

Optimizes AI agent prompts for clarity, token efficiency, and actionability. Refines SKILL.md files and agent definitions to enhance AI performance.

Builds a retrieval-optimized knowledge layer over agent documentation, mapping contexts to knowledge chunks and enhancing information density for efficient agent consumption.

Enables multi-agent coordination for complex workflows, including parallel code reviews and dependency-based pipelines using Claude's TeammateTool and Task system.

Crafts correct SQL queries for the Lunar PostgreSQL data model to analyze components, checks, policies, domains, and PRs.

Configures and standardizes agent behavior definitions in AGENTS.md for coding agents, ensuring clarity and repo-specific context in multi-agent systems.

Provides Supabase-recommended best practices for optimizing Postgres queries, schema design, and database configurations to enhance performance.

Orchestrates parallel AI coding agents (Claude Code, Codex, OpenCode) in isolated Git worktree and tmux environments with real-time status monitoring.

Configures vector embeddings and semantic search for document collections to enable AI-powered similarity search and RAG knowledge base preparation.

Updates and maintains AI agent memory files including context, progress, and decision logs for seamless operation.