Luke Nam

CS PhD @ Vanderbilt

San Francisco Bay Area

About

Hi, my name is Luke Nam, a first-year Ph.D. student at Vanderbilt University studying computer science with a research interest in neural network verification and formal methods. Previously, I was a student at Duke University, and graduated with a BS in Computer Science in May 2025. I'm proficient in Python and C/C++, and dedicated to working collaboratively in research labs. I watch college football and basketball in my free time, and a former clarinet in the Duke University Marching and Pep Band for all four years. You can find more of my work at my personal website as listed here: https://lukelike1001.github.io/

Experience

  • Research Intern at Oak Ridge National Laboratory
    May 2026 - Present · 3 mos

    Researching formal verification of drones against adversarial attacks. Developing the Drone Adversarial Robustness Toolbox (DART) open-source library under the mentorship of Dr. Amir Sadovnik.

  • Undergraduate Teaching Assistant at Duke University
    Jan 2025 - May 2025 · 5 mos

    - Recipient of the Most Outstanding UTA Award for Duke CS370 (Intro to Artificial Intelligence) - Led weekly discussion sections, hosted office hours, answered online forum questions, and graded both assignments and exams

  • NREIP Research Intern at U.S. Naval Research Laboratory
    May 2024 - Feb 2025 · 10 mos

    - Undergraduate Intern at the Navy Center for Applied Research in Artificial Intelligence (NCARAI) - Co-authored a paper for enhancing naval systems with LLMs accepted to AAAI 2025 (linked below) - Improved naval travel itinerary algorithm accuracy from 0.6% to 90% using Large Language Models (LLMs) - Utilized ChatGPT-4o to develop basic military courses of action (COAs) peaking at 99% accuracy - Developed human feedback interfaces for military LLM applications - Constructed a speech-to-text (STT) model pipeline for aerial combat planning - Improved reasoning and self-reflection capabilities for state-of-the-art LLMs using ReAct

  • Undergraduate Student Researcher at Duke University
    May 2023 - Dec 2023 · 8 mos

    - Fine-tuned a time series model to predict earthquake magnitude, time, and location - Collected >7.97 million earthquakes from 23 observatories using AWS Lambda, S3, EC2, EKS, and BeautifulSoup