3. Data, Science & AI

Optimizes AI model interactions through advanced prompt design, enhancing response quality and applicability in LLM-based systems.

Provides quantitative trading strategy development for DeFi and crypto derivatives, including risk management, backtesting, and performance metrics analysis.

Manages the lifecycle of AI agent skills including creation, testing, and organization, aligned with Claude Code documentation.

Generates dynamic prompts using templates and patterns for optimized LLM interactions, supporting RAG, fine-tuning, and agent workflows.

Dispatches multiple independent AI agents to concurrently investigate and resolve unrelated system failures without shared dependencies.

Builds, audits, and optimizes LangGraph.js agents for stateful AI workflows in TypeScript/JavaScript.

Orchestrates multiple AI coding agents to work concurrently on GitHub issues using git worktrees, enabling parallel development of related features.

Validates machine learning workflows to prevent common antipatterns including data leakage and evaluation errors during training and deployment.

Generates SQL migration scripts for Supabase and PostgreSQL database schema changes, including table creation, column modifications, and index additions.

Integrates existing SQL into data stack layers (ODS/DIM/DWD/ADS) by adding tables, configuring data sources, and generating migration files.

Assists in writing efficient SQL queries, optimizing database performance, and conducting data analysis for actionable insights.

Build production ChatKit chatbots integrating OpenAI Agents SDK, MCP tools, and custom backends for real-time task execution and user isolation.

Sets up agent guidance, rules, and onboarding documentation (AGENTS.md) to ensure AI agents comply with repository standards and external skills.

Comprehensive guide for constructing multi-agent AI systems using openai-agents SDK with live documentation via Context7.

Designs systematic frameworks to select optimal AI tools based on task requirements, codebase size, and reasoning depth for efficient resource allocation.

Provides pre-usage validation to optimize expensive LLM resource consumption, preventing unnecessary calls and reducing costs.

Expert guidance for managing Qdrant Cloud collections, optimizing vector schemas for RAG, and executing semantic search queries in Python.

Orchestrates parallel Claude AI agent execution, preventing file conflicts and enabling real-time status tracking via command interface.

Optimizes PostgreSQL query performance, scales database design, and debugs performance bottlenecks for efficient data handling.

Creates reusable prompt templates for recurring AI tasks, capturing domain-specific patterns to eliminate repetitive prompt engineering.

Enables production hybrid search using PGVector HNSW with BM25 and Reciprocal Rank Fusion for semantic + keyword retrieval in vector databases.

Provides PostgreSQL query optimization patterns and migration review guidelines for efficient database management and debugging.

Creates structured LLM prompts with seven elements (Objective, Constraints, Outcomes) to clarify AI agent goals and decision parameters.

Persistent Python REPL server that maintains state and avoids reloading heavy ML packages for efficient iterative model testing and development.