3. Data, Science & AI

End-to-end title processing pipeline handling data collection, embedding, comparison, and formatting for comprehensive title analysis.

Generates GA4 analytics reports for KStoryBridge dashboard, focusing on user acquisition, growth metrics, and real-time monitoring through funnel analysis and traffic source breakdowns.

Defines execution rules for AI agents including TDD, QA, context management, and completion reports for all development tasks.

Collects and processes title metrics from Korean webtoon/webnovel platforms and fan engagement sources for data-driven intelligence.

Transforms documents into structured chunks with metadata for efficient Retrieval-Augmented Generation (RAG) systems.

AI-powered documentation lookup for the Awareness project using RAG, automatically identifying and improving technical documentation gaps.

Optimizes and standardizes data interactions with Supabase, implementing data engineering best practices for efficient database operations.

Standardizes and controls Gemini AI API calls by separating prompts from code, improving maintainability and robustness.

Transforms simple requests into optimized prompts using Chain of Thought, Persona, and Constraints to maximize LLM output quality.

Analyzes code to identify AI agent implementation opportunities and visualizes options through flowcharts for clear technical explanation.

Generates AI-powered images via Gemini API for design assets including landing pages, product visuals, and placeholders, saving to repository.

Analyzes and fixes ineffective subproblems in SciCode benchmark, classifying issues as MAJOR/MINOR and applying minimal fixes to MAJOR issues.

Reviews and validates scientific code benchmark problems for quality, correctness, and test case accuracy.

Generates and executes Snowflake SQL queries from natural language input, outputting results in CSV format.

Enables semantic search within CasareRPA codebase using ChromaDB and MCP server to locate code by intent without exact names.

Enables execution of SQL queries, table inspection, and database administration for the TMI PostgreSQL system.

Creates temporary subagents with isolated context and task-specific boundaries to enhance AI workflow modularity and reduce interference.

Optimizes database performance by validating schemas, identifying missing indexes, and auditing Prisma ORM usage patterns.

Comprehensive tool for spreadsheet creation, editing, analysis, and visualization with formula support and intelligent data import capabilities.

Executes reproducible Python scripts for data analysis, statistical modeling, and visualization to explore datasets and build predictive models.

Queries remote Parquet files via HTTP using DuckDB's httpfs with column pruning and row filtering for efficient crypto/trading analytics.

Comprehensive guide for developing AI agents with SDK v6, covering tool definitions, multi-step workflows, and result extraction patterns.

Integrates AI/ML models, RAG systems, and vector databases into web applications with API setup, prompt engineering, and model deployment.

Reference guide for AI-based protein sequence design using LigandMPNN and ProteinMPNN, including amino acid biasing, fixed residues, and membrane protein applications.