Soungmin (Min) Lee

about

georgia tech

who

Portrait of Soungmin (Min) Lee — fig. me · georgia tech

language, speech, and the software around them.

Hello, I’m Soungmin (Min) Lee, a B.S./M.S. student in Computer Science at Georgia Tech, specializing in NLP and web development. Currently focused on NLP research and integrating LLM-powered features into web/mobile services, I enjoy building software solutions that optimize critical processes and solve complex problems.

exp. 2027: m.s. computer science
machine learning · georgia tech
2026: b.s. computer science
info-internetworks / intelligence · georgia tech
based: atlanta, ga

research

03 papers

The robot soldier from Castle in the Sky on the Ghibli Museum rooftop at dusk — fig. robot soldier, castle in the sky ghibli museum, mitaka photo by me

latest 2026.06 icml 2026 workshop on machine learning for audio

Towards Truly Multilingual ASR: Generalizing Code-Switching ASR to Unseen Language Pairs

Gio Paik, Hyunseo Shin, Soungmin Lee

Investigates whether code-switching ASR capabilities learned from limited seen language pairs can generalize to unseen language pairs using model merging and domain generalization methods.

read paper paper arxiv hugging face

2026.01 findings of eacl 2026

HiKE: Hierarchical Evaluation Framework for Korean-English Code-Switching Speech Recognition

Gio Paik, Yongbeom Kim, Soungmin Lee, Sangmin Ahn, Chanwoo Kim

paper arxiv code dataset
2025.12 neurips 2025

Words That Unite The World: A Unified Framework for Deciphering Central Bank Communications Globally

Agam Shah, Siddhant Sukhani, Huzaifa Pardawala, Saketh Budideti†, Riya Bhadani†, Rudra Gopal†, Siddhartha Somani†, Michael Galarnyk†, Soungmin Lee†, et al.

paper code dataset poster website

all research →

work

since 2021

Adrián Villar Rojas' installation The Language of the Enemy at Art Sonje Center: scorched metal, cables and roots hanging in a gallery — fig. the language of the enemy adrián villar rojas, art sonje center photo by me

roles: 05
teams: 04
since: 2021

2026.06 — 2026.08 Software Development Engineer Intern Amazon Web Services ebs · vmds
seattle, wa
- Returned to the Elastic Block Store VMDS team.
2025.08 — 2025.11 Software Development Engineer Intern Amazon Web Services ebs · vmds
seattle, wa
- Elastic Block Store VMDS team.
2024.01 — 2025.08 Undergraduate Researcher Georgia Tech VIP Program fintech lab
atlanta, ga
- Core contributing author of the World Central Banks (WCB) dataset paper, published at NeurIPS 2025, introducing the largest monetary policy corpus to date and benchmarking state-of-the-art language models for stance, temporal, and uncertainty detection.
- Implemented pipelines and filtration techniques to counter intentional mixed tone of central banks of countries such as the USA, South Korea, Turkey, etc.
- Engineered a new dataset format which drastically improves classification accuracy of hawkish/dovish stances of the Fed using pre-trained language models.
- Analyzed NLP models such as StockNet from mid-2010s using Python and TensorFlow, evaluating if the models hold accuracy with modern data.
2024.07 — 2024.08 Software Engineer Intern LG Innotek seoul, south korea
- Developed a trilateration localization system using BLE beacons from NXP, implemented in C and packaged with CMake, achieving 9 yard accuracy for Bluetooth device positioning.
- Applied and trained the ECA-ResNet model using Python and TensorFlow on a RSSI localization system, improving vehicle tracking accuracy by 15% and identifying key error factors for further optimization.
- Implemented a customized particle filter for UWB data, reducing location measurement error by 55% within an 10 yard range, enhancing mobile digital key functionality.
2021.09 — 2023.03 Software Developer, Sergeant Republic of Korea Air Force 지능정보체계관리단
gyeryong, south korea
- Mandatory military service.
- Affiliated unit: Intelligent Information Systems Management Group (지능정보체계관리단).
- Java, JavaScript, Python, ReactJS, ExpressJS, Spring Boot, SQL.