Bachelor of Arts (B.A.) — Data Science, Integrative Biology
Berkeley Certificate in Design Innovation

• Coursework: Machine Learning, Data Engineering, Data Structures, Principles & Techniques of Data Science, Probability for Data Science, Optimization Modeling, Structure and Interpretation of Computer Programs, Multivariable Calculus, Linear Algebra & Differential Equations, General Biology, Organic Chemistry

Abbasi Lab, UCSF, San Francisco, CA

• Developing an LLM-based EHR summarization pipeline at UCSF, auditing retrieval over a large-scale synthetic EHR database (2M+ clinical records) and designing a claim-level grounding evaluation framework to measure and reduce hallucination in clinician-facing outputs.

SmartMind AI, Seoul, Republic of Korea

• Initiated and led a natural-language-to-SQL pipeline on a live enterprise ERP system, mapping the 6 core tables of 20 required for coverage and replacing single-run pass/fail testing with a multi-trial measurement design over a fixed golden set of query scenarios, reporting a measured pass rate with run-to-run variance.
• Traced silently failing queries back to the Neo4j ontology (ApeRAG/Qurify) that supplied the model its schema and business knowledge, auditing entity-merge accuracy and similarity thresholds to surface missing and duplicate nodes, and corrected errors in the reference answer key by reconciling output against source financial statements.

Asan Medical Center (AMC), Seoul, Republic of Korea

• Ran large-scale exploratory analysis in Python and Jupyter for the VLM3D Multi-Abnormality Classification Challenge, identifying the data patterns that informed feature selection and a strategy placing 7th among multi-institutional teams.
• Developed and fine-tuned PyTorch deep learning models (3D U-Net, ResNet3D) for medical imaging and a multimodal deep survival model for hepatocellular carcinoma, using SHAP and Grad-CAM to validate feature importance.

Invention Corps of Berkeley; Berkeley Consulting Collaborative

• Mozilla Firefox: structured a five-theme survey instrument with a 12-person team and analyzed responses into the feature requirements behind a shipped React and Vite extension tracking 50+ concurrent tabs by memory and CPU usage.
• Braven: cleaned and visualized survey data with an 8-person team, then re-segmented by class year, producing four cohort personas that drove the wireframes and 6 production-ready mobile screens delivered in Figma.
• Replate: analyzing Replate's large-scale user data to design and build a website dashboard used by both donors and the company.

US 8th Army, 2ID HHBN

• Led a platoon of 40+ soldiers as a bilingual Korean-English Platoon Leader, translating and delivering ACFT training and basic military coursework, resulting in a 96% pass rate across the entire platoon.
• Awarded the US Army Certificate of Achievement (Mar. 2024) for outstanding leadership and bilingual communication.

Cal Galacticos — UC Berkeley Intramural Competitive Soccer

• Captained the largest Korean soccer team in Berkeley, recruiting new members, managing events and social media, and running team logistics across competitive seasons.