Yanggee Kim wheresmadog

Yanggee Kim

LLM / AI Agent Engineer (Python) · 고려대학교 기초과학연구원

Designing, building, and operating Python-based LLM/AI Agent services — RAG, tool calling, and permission-aware guardrails.

Build RAG and tool-calling agent workflows, integrate external APIs, and apply guardrails so answers stay within access rights
Standardize agent project structure and deployment so services run reliably in production
Comfortable with Linux/Docker, Git-based collaboration, code review, testing, and release
Background in ML model training and deployment (including medical imaging)

Python · RAG · LangGraph · Tool / Function Calling · LLM APIs · Docker · Linux · PyTorch · Git

University messenger AI Agent — development and operations

RAG Q&A agent plus tool-calling workflows for attendance, quizzes, and related campus tasks
Adapters to external systems/APIs with guardrails that limit answers by viewer permissions
Reusable project layout, prompts/tools, and deployment setup for ongoing service operation

Cardiac ultrasound AI and on-device performance

Built data storage and training environments; standardized multi-stage training and model deployment for the team
Shipped a review UI so predictions could be checked and iterated quickly
Cut peak memory ~19% on Windows on-device AI via caching; improved latency by fixing bottlenecks

2023 인공지능 그랜드 챌린지 2단계 입상
2022 한국어AI경진대회 우수상
Paper: Yanggee Kim, Hanyoung Kim, Donghun Lee. (2023). Aspect-based Dense Passage Retrieval. 한국정보과학회 학술발표논문집
M.S. Applied Mathematics, Korea University (고려대학교 일반대학원 수학과)