An access-control check that only lived in the UI, and what it took to trace it to the database, plus the harder question it raised about how much access an AI agent should have.

UKC FIRE 2026: When AI Becomes Part of the Security Boundary

Blog Redesign : The Role That Wasn’t in the Room

Supply Chain Security: When a Stopped Crash Hides a Live Defect

Coordinated a 3-day traditional craft workshop at NMAA, facilitating cultural exchange and securing a meaningful book endorsement from prominent heritage authors.

K-Heritage Craft Workshop: Cultural Community Outreach

Served as a bilingual event volunteer at the Korean Cultural Center, bridging heritages and gaining deep insights into cultural diversity through the lens of Korean literature

K-Culture Event: Bridging Borders Through Korean Literature

Explores the methodology of training Large Language Models to identify and analyze security flaws within source code.

LLM Security: 2. How LLMs Learn to See Vulnerabilities

An introductory overview of the comprehensive research series exploring vulnerabilities, alignment breaks, and defensive tactics in LLM environments.

LLM Security: 0. About This Series


Assesses the dual-use nature of LLMs, focusing on their economic viability and capabilities in executing autonomous cyber attacks.

LLM Security: 5. LLMs on the Offensive: Capability, Economics, and Autonomous Attack

Investigates techniques and prompt engineering strategies for guiding LLMs to generate secure, robust, and vulnerability-free code.

LLM Security: 3. Teaching Models to Write Safer Code

Analyzes unique attack vectors targeting autonomous LLM agents and evaluates proactive defense frameworks to secure them.

LLM Security: 4. Attacking and Defending LLM Agents

Examines how adversarial prompt injections bypass safety alignments and manipulate Large Language Model outputs.

LLM Security: 1. When Alignment Breaks Adversarial Attacks and Prompt Injection

Shared insights and personal growth gained from serving as a primary organizer for the IMPACT 2026 conference.

IMPACT 2026: First Organizing Experience

Mentored aspiring developers at an AI coding camp, sharing technical insights and fostering the next generation of engineers.

AI Coding Camp: Tech Mentoring

Facilitated a large-scale community cultural event by efficiently managing check-in operations and logistics.

KIMJANG DC: Community Event Operations

Presented research on an Automated Build System integrated with LLMs, exploring the intersection of AI and DevOps.

KOCSEA 2025: LLM-Based Automated Build System

Supported international cultural exchange efforts through volunteer coordination at a major public museum event.

NMAA Chuseok Event: Cultural Exchange Coordination

Executed a comprehensive security project spanning from web application exploitation to system-level Capture The Flag (CTF) challenges.

Pentest Project: Web App to System CTF


Documented the migration process and code refactoring required to update the blog system to the Notion API v2025-09-03 version.

Blog: Notion API v2025-09-03 Migration🚀

Successfully completed a first U.S. half-marathon, showcasing personal discipline, endurance, and goal-setting.

DC Half Marathon: First U.S. Solo Race

Analyzed the root causes of packet drops and performance degradation when implementing TLS security layers on real-time operating systems.

RTOS meets TLS: Analyzing Packet Drops

Investigated the trade-offs and integration strategies between real-time processing constraints and TLS cryptographic overhead in IoT networks.

RTOS meets TLS: Performance vs Security

Provided critical communication support between press and AV teams during a high-profile international diplomatic summit.

Korea-US Summit: Press & AV Team Coordination


Delivered a career flash talk focused on professional resilience and maintaining technical integrity in the industry.

UKC FIRE 2025: Staying Firm Without Snapping

Developed an automated blogging workflow that synchronizes GitHub open-source repositories with Notion and auto-deploys via Vercel.