Role
AI Researcher
// IDENT — 0x01
EUISUH JEONG
aka
[euisuh]
02 · about
Engineer by training, researcher by drift.
A short dossier on origins, trajectory, and the cultures that shaped the way I read systems.
callsigneuisuh
bornSeoul, KR
passports4 countries
fluentKO · EN
workingHI · AR (some)
stackPython · Torch
edgeFastAPI · Celery
infraRedis · Docker
I'm an AI researcher and software engineer currently serving as a Staff Sergeant with the Republic of Korea Air Force, where I build computer-vision systems for runway integrity.
Before the service, I helped found AIxamine at QCRI — a platform that stress-tests language models against safety benchmarks. I'm a Carnegie Mellon CS '22 grad, with a minor in mathematical studies.
I've been moving since I was three. Seoul, then a small town in the US, then back to Seoul, then India for secondary, then Qatar and the US for undergrad, then Qatar for work, then home again. Cultures stack, like middleware. The interesting work happens in the seams.
// trajectory.log
● Seoul, KR '98 ── ● —, US '01 ── ● Seoul, KR '04 ── ● —, IN '10 ── ● Doha, QA '18 ── ● Pittsburgh, US '20 ── ● Doha, QA '22 ── ● Seoul, KR '25→
kindergarten · elementary · secondary · cmu doha · cmu pittsburgh · qcri · rokaf · city, country — drop the kindergarten + secondary city names in when you'd like
03 · now
What I'm doing this week.
Updated automatically.
Runway integrity, end-to-end.
Building and maintaining the AI backbone for a runway evaluation system at an active ROKAF airbase: pavement crack and defect detection, automated PCI scoring, image pipeline triage. Currently optimizing inference latency on the labeling-tool side and re-balancing the segmentation head against winter-condition data.
- Re-training the defect classifier on a fresh 12k batch
- Reading: Vision-Language Models for Industrial Inspection
- Pulling Korean military service to a close in Q4 — looking for what comes next
04 · experience
Six years, three time zones.
Full record on LinkedIn. Highlights below.
2025 — present
AI Researcher · Staff Sergeant · Squad Leader
Republic of Korea Air Force / AI-Based Technology Team
Led squad developing a deep-learning system for automated runway pavement evaluation. Directed reconstruction of a high-quality dataset (sourced from low-quality captures, valued north of $500K) and drove a 38% relative accuracy gain — from ~62% to >84% — through feature engineering, postprocessing, and targeted data curation against model bias. Full project lifecycle as technical lead; authored the project paper.
2023 — 2025
Research Assistant
Qatar Computing Research Institute (QCRI)
Co-developed aiXamine — a black-box LLM safety evaluation platform with 40+ benchmarks across 8 security dimensions. Built the modular reporting + visualization architecture; evaluated 50+ models across 2K+ exams, surfacing vulnerabilities in GPT-4o, Grok-3, and Gemini 2.0. Also investigated backdoor Trojan attacks on code-focused LLMs (finetuning + susceptibility testing).
2022 — 2023
Junior Software Engineer
KARTY · Spend, Save, and Manage
Built a multi-channel notification system (SMS, email, push) for the consumer fintech app. Migrated payment processing to a compliant platform under regulatory scrutiny. Designed and shipped a Clubhouse-style waiting list + lottery system tied to FIFA World Cup Qatar 2022.
2021 — 2022
Teaching Assistant · 11-785 Deep Learning (PhD-level)
Carnegie Mellon University · Pittsburgh
Planned and delivered lectures, recitations, and assignments to 350+ students in CMU's flagship deep-learning course. Mentored research projects and guided exploration of novel directions.
2018 — 2022
B.S. Computer Science · Minor, Mathematical Studies · University Honors
Carnegie Mellon University
Split between the Doha and Pittsburgh campuses. Coursework concentrated in systems, machine learning, and applied math.
05 · projects
Things I built that went live.
A non-exhaustive list. The ones I can talk about are below; ask for the rest.
P-01 CV / INFRA
Live · ROKAF
Runway Evaluation System
Tech lead · data + model + backend
Detects cracks and surface defects on airbase runways and computes PCI scores from drone-collected pavement imagery. I instrumented the data-labeling pipeline, commanded the labeling team, trained the segmentation + classification backbone, and built the async FastAPI + Celery + Redis backend that ties capture, queue, inference, and reporting together. Currently in operational use.
PyTorchFastAPICeleryRedisPostgresDocker
P-02 LLM / EVAL
Live · public
AIxamine
Founding member · benchmark harness
A safety-evaluation platform for language models. Run a model through a battery of bias, robustness, and jailbreak benchmarks; get an honest scorecard back. One of the founding members; co-author on the accompanying paper.
PythonLLMEvalBenchmarks
P-03 LLM / SECURITY
Thesis · 2024
LLM Code Poisoning & Vulnerability Induction
Master's thesis · sole author
Dissertation exploring data-poisoning attacks that coerce code-LLMs into emitting vulnerable source. Designed and tested stealthy trigger-based backdoors; fine-tuned code LLMs on poisoned corpora to measure susceptibility; analyzed model limitations in vulnerability detection. Companion work to the AIxamine research direction.
LLMSecurityBackdoorsPyTorch
06 · publications & talks
Papers and conference work.
Newest first.
arXiv · 2025
AIxamine: A Comprehensive Safety Evaluation Platform for Large Language Models
… · E. Jeong · … (see paper for full author list)
slot · TBD
Add additional talks or publications here
authors · venue
07 · writing
Notes & long-form.
Paste real posts in
operator.html when you have them. Until then, these are scaffolds.// 4 placeholder posts — open one to see the reader, then replace bodies in the
POSTS array08 · places
Six cities, one stack.
Drop photos in later — these cells are sized correctly and will swap cleanly.
1998
SEOUL · born here
2001
USA · kindergarten
2004
SEOUL · elementary
2010
INDIA · secondary
2018
DOHA · CMU-Q
2020
PITTSBURGH · CMU main
2022
DOHA · QCRI · AIxamine
2025
SEOUL · service
Get in touch.
Open to research collaborators, post-service roles, and the occasional good email. Fastest reply on LinkedIn.