- Conduct in-depth R&D on digital human embedding technologies, focusing on enhancing realism in visual representations- Engineer AI-driven applications, from initial development through deployment, supporting innovative digital media and visualization projects.- Design and optimize data processing pipelines to streamline the integration of datasets into AI workflows- Leverage the latest advancements in vision-related research and technology to drive improvements in digital human and visual AI applications.

- Collaborated in developing & fine-tuning LLM models to create a valuation algorithm for automating investment decisions, utilizing Few-Shot In-Context Learning for higher accuracy.- Helped Develop a finance-oriented Multimodal Retrieval-Augmented Generation (RAG) system that integrated over 80,000 financial reports, achieving a reduction in hallucinations by over 60%.- Designed evaluation pipelines to validate LLM-generated investment insights, improving precision and recall metrics.

MATH 3640: Probability (Fall 2023)MATH 2410: Methods of Linear Algebra (Spring 2023)

- Created Data Quality management rules for evolving data landscapes and domains in SQL database - Optimized incompatibility between old/new server database; improved 40,000+ misleading sales data

- Analyzed B2B supply chain management data, providing insights through Integrated Business Programming Language (IBPL) to optimize client operations.- Developed predictive models within the platform, delivering masked data to facilitate informed decision-making for clients.- Generated equations to estimate international logistics costs and total profits, enhancing the client's financial forecasting capabilities.

- Performed SQL analysis on user-retention rate from CRM data; enrolled 15,000+ new/returning customers - Data Pipeline Construction Project: re-structured data base in 3 levels: DataSource, DataBase, DataMart- Reduced commercial cost by 30%+ by obstructing inefficient ad-channel; created Dashboard for metrics monitoring- SQL Query Education for Product/Brand Managers: instructed how to write SQL queries for simple analysis by using DataMart to the managers of product teams - LTV Forecasting Model Development Project: Anticipated the first and second repurchase rate of each three brands in company by deploying curve-fitting method and using SQL and Python through Zupyter. Divided the customers into different cohorts based on the first purchase months and four different order dynamics composed of one-time purchase and subscription. Computed the forecasted retention rate and LTV in the average accuracy of 92%.

• Led the Artillery Forward Observer Special Forces Squad in 18 joint military exercises, providing corps-level tactical communication/battlefield fire direction control in the Southern FEBA (Forward Edge of Battle Area) in Capital Corps of Republic of Korea Army

Instructed international students eager to achieve deeper understanding of Mathematics including Linear Algebra, Multivariable Calculus, AMC 10