I'm a Data Scientist and AI Engineer with experience in applied machine learning, data analysis, and AI system development.

My work has focused on turning complex domain data into reliable, service-ready systems.
I contributed to a public-sector project for agricultural reservoir water-level forecasting and disaster prevention, working across data collection, preprocessing, model development, evaluation, and database-backed result management.

I also have experience in document intelligence and search-oriented data structuring, including PDF/OCR post-processing, caption extraction, related sentence matching, and structured data generation for retrieval.

Recently, I have been exploring practical AI systems around AI Agents, Graph RAG, ontology-based knowledge modeling, and LLM-driven analysis workflows through personal projects.

My main interest is building AI systems that are not only accurate, but also reproducible, explainable, and useful in real operational environments.

• Time-series forecasting and predictive modeling
• Data preprocessing, anomaly handling, and feature engineering
• ML model development and evaluation with Python and PyTorch
• Database-backed data workflows with MySQL and MariaDB
• Search-oriented document structuring using OCR, TF-IDF, and parsing logic
• Docker-based development and deployment environments
• LLM-based analysis workflows and structured output generation
• Knowledge-driven AI systems using Graph RAG, ontology, and agentic workflows

Built and improved a time-series forecasting workflow for agricultural reservoir water-level prediction.

• Unified reservoir, canal, and weather data into hourly time-series datasets
• Designed preprocessing rules for missing values, outliers, abnormal spikes, and inconsistent timestamps
• Compared forecasting models including NLinear and DLinear
• Improved model performance through architecture adjustment and loss function experimentation
• Managed processed data and prediction results through database-backed workflows

Worked on structuring research paper data for better search and retrieval.

• Processed OCR and layout detection results from PDF documents
• Extracted table and figure captions using rule-based proximity logic
• Connected captions with related sentences and keywords using TF-IDF-based scoring
• Generated structured JSON/RDB-ready data bundles for downstream search and storage

Personal project focused on AI search visibility and LLM-driven website analysis.

• Crawls and analyzes website SEO/AEO/GEO signals
• Tracks brand mentions, official links, and citation presence in LLM responses
• Uses structured outputs to generate improvement actions
• Explores practical LLM workflows for analysis, scoring, and report automation

Personal research and development around knowledge-driven AI applications.

• Exploring Graph RAG structures for retrieval beyond plain vector search
• Studying ontology-based data modeling for better entity and relationship representation
• Building small AI agent workflows that combine retrieval, reasoning, and structured actions
• Interested in connecting RDB, vector search, graph databases, and LLM agents into practical systems