Overview
Work History
Education
Skills
Awards
Leadership & Teaching Experience
Timeline
Generic

YOUNGMIN (YOUNG) KO

Evanston,USA

Overview

4
4
years of professional experience

Work History

World Model of Large Language Models

Northwestern & Yale Universities
01.2025 - 05.2025
  • Co-authored LAMP: Extracting Locally Linear Decision Surfaces from LLM World Models (under review)
  • Fits locally linear surrogates to LLM self-explanations, revealing how stated factors impact decisions, achieving R² values (e.g., 0.42±0.03 for GPT-4.1-mini) comparable to LIME with enhanced feature interpretability.
  • Validated across sentiment, safety, and clinical tasks (e.g., gastroenterology expert agreement of 0.697 vs. 0.635 inter-expert baseline) showing LAMP outputs align with human judgment.

Diagnostic Reasoning With Large Language Models

Northwestern & Yale Universities
11.2024 - 03.2025
  • Presented poster Automated Prompt Optimization Strategy Improves Large Language Model Diagnostic Accuracy for Complex Clinical Cases in Gastroenterology and Hepatology at Digestive Disease Week 2025
  • Engineered an evolutionary prompt-optimization framework, boosting GPT-4.1-nano top-1 diagnostic accuracy on 262 real-world gastroenterology cases from 59.3% to 74.6% (15.3 pp increase).

AI Safety in Evidence-Based Medicine

Northwestern & Yale Universities
05.2023 - 09.2024
  • Co-authored Expert-of-Experts Verification and Alignment (EVA) Framework for LLM Safety in Gastroenterology – accepted in npj Digital Medicine (Nature Portfolio journal)
  • Benchmarked 9 open-source & proprietary LLMs (GPT-3.5/4/4o/1, Claude-3-Opus, Mixtral 7B, Llama-2 7B/13B/70B) across 27 baseline, RAG, and LoRA-SFT configurations, identifying a GPT-4o + RAG setup that exceeded board-certified accuracy by 12 pp
  • Trained a reward model (OPT-350M) on 7k expert-graded answers, raising automated grading precision from 80.2 % to 87.9 % and enhancing end-to-end answer accuracy by 8.4 % via rejection sampling

REU on Stochastic Processes and Financial Time Series

Yeshiva University
06.2021 - 08.2021
  • Detected anomalies in equities via Persistence Landscape by embedding daily returns time series data with Takens Embedding Theorem with gudhi package
  • Presented the work Capturing Patterns in Equity Price Change via TDA under the guidance of Professor Marian Gidea in Yeshiva University Mathematical Physics seminar: REU student presentations

Teaching Assistant

Northwestern University
09.2023 - 12.2023
  • STAT436: Reinforcement Learning, STAT362: Advanced ML for Data Science, STAT320: Statistical Theory&Methods, STAT304: Data Structures and Algorithms for Data Science

Team Lead

Northwestern University, English Language Program
06.2023 - 08.2023
  • Led 8 peer mentors as a team lead in supporting 67 international graduate students, organizing weekly check-ins, delegating tasks, and providing resources to enhance discussions and engagement

Workshop Teaching Assistant

Northwestern University, Institute for Policy Research
06.2024 - 08.2024
  • Provided instructional and logistical support for two intensive summer research training institutes for researchers:
  • NSF-Funded Summer Research Training Institute: Improving Evaluations of R&D in STEM Education (July 8-12, 2024): Prepared instructional materials, led lab sessions, and assisted in breakout groups focusing on evaluation methodologies in STEM education research.
  • IES-Funded Workshop: Randomized Control Trials (2024 RCT Summer Institute) (July 15-25, 2024): Mentored participant groups through daily group projects and analysis labs covering experimental design, statistical modeling, power analysis, and validity considerations in RCTs.

Group Leader

Northwestern University,
09.2024 - 11.2024
  • Curated events as a group leader for 10+ participants per event at Chicago’s iconic venues to immerse international students in local culture and boost English confidence

Education

PhD Candidate - Statistics

Northwestern University
Evanston, IL
06.2027

Bachelor of Science - Mathematics

Haverford College
Haverford, PA
05.2022

Skills

  • Programming: Python, R
  • Frameworks: PyTorch, Langchain, Pandas, NumPy
  • Tools: Git
  • Machine Learning: Reinforcement Learning, NLP, Large Language Models, Reward Modeling

Awards

Haverford KINSC/Velay Summer Scholar Fellowship May 2021, Haverford First Year Math Prize Exam, Single Winner March 2017

Leadership & Teaching Experience

  • Teaching Assistant: STAT346 (Reinforcement Learning), STAT362 (Advanced ML for Data Science), STAT320 (Statistical Theory&Methods), STAT304 (Data Structures and Algorithms for Data Science)
  • Workshop Assistant: Improving Evaluations of R&D in STEM Education (NSF-Funded Summer Research Training Institute), 2024 Randomized Control Trials Summer Institute (IES-Funded Workshop)
  • Group Leader: English Development through Guided Experience (EDGE), Northwestern University
  • Team Lead: English Language Program (ELP), Northwestern University

Timeline

World Model of Large Language Models

Northwestern & Yale Universities
01.2025 - 05.2025

Diagnostic Reasoning With Large Language Models

Northwestern & Yale Universities
11.2024 - 03.2025

Group Leader

Northwestern University,
09.2024 - 11.2024

Workshop Teaching Assistant

Northwestern University, Institute for Policy Research
06.2024 - 08.2024

Teaching Assistant

Northwestern University
09.2023 - 12.2023

Team Lead

Northwestern University, English Language Program
06.2023 - 08.2023

AI Safety in Evidence-Based Medicine

Northwestern & Yale Universities
05.2023 - 09.2024

REU on Stochastic Processes and Financial Time Series

Yeshiva University
06.2021 - 08.2021

Bachelor of Science - Mathematics

Haverford College

PhD Candidate - Statistics

Northwestern University
YOUNGMIN (YOUNG) KO