3. Data, Science & AI

Designs efficient SQL and NoSQL database schemas with normalization, indexing, and migration strategies for optimal data organization and performance.

Optimizes AI agent prompts and RAG systems by reducing token usage while preserving response quality through strategic context positioning.

Provides advanced RAG patterns including Self-RAG, Corrective-RAG, and knowledge graphs for adaptive retrieval and query rewriting in agentic systems.

Provides reusable patterns for cross-encoder reranking and LLM-based relevance scoring to enhance search precision in RAG pipelines.

Provides LLM observability for tracing, evaluation, prompt management, and cost tracking in AI applications.

Provides production-ready hybrid search using PGVector HNSW and BM25 with Reciprocal Rank Fusion for optimized semantic and keyword retrieval.

Quantitative risk analysis for trading strategies using Kelly criterion, tail risk metrics, and Monte Carlo simulations to optimize position sizing and capital allocation.

Performs comprehensive health checks on trading strategies by analyzing performance metrics, stress testing, and identifying risk indicators from historical data.

Validates trading strategy robustness by testing optimized parameters on out-of-sample data through walk-forward analysis.

Analyzes trading backtest data to identify performance patterns across parameters such as time of day, DTE, and market conditions for strategy optimization.

Analyzes trading strategy portfolios to evaluate correlation, diversification effects, and combined performance for optimized strategy selection.

Provides data validation and normalization utilities using Nette Schema for configuration and API input structures.

Enables SQL-based querying and analysis of Claude Code history combined with Git commit data for productivity and coding pattern insights.

Dispatches multiple Claude AI agents in parallel to investigate and resolve independent issues simultaneously, improving efficiency.

Manages user preferences and auto-learns from session corrections to personalize bilateral modeling interactions.

Guides SQL query writing, schema analysis, and migration design specifically for SQLite database operations.

Guides creation and editing of Observable Notebooks 2.0 for interactive data visualization and exploration in data science projects.

Enables real-time Vietnamese AI interaction with action-first execution, minimizing explanations while prioritizing task completion.

Optimizes prompt design for Anthropic's Claude AI models using proven engineering methodologies to enhance output quality and reliability.

Generates and refines prompts for LLMs using advanced techniques like chain-of-thought and few-shot prompting to enhance output quality.

Analyzes project requirements to recommend optimal Anthropic-based AI architecture, including Skills, Agents, Prompts, and SDK primitives.

Routes user requests to specific LLM providers (OpenAI, Grok, Groq, DeepSeek, OpenRouter) using SwiftOpenAI-CLI for agent mode.

Expertly optimizes SQL queries, analyzes performance, and designs indexes for PostgreSQL, MySQL, and MongoDB to resolve slow query issues.

Provides plain-language summaries of PubMed scientific papers for medical research, clinical trials, and biomedical topics.