저는 배운 내용을 외운 지식으로 끝내지 않고, AI와 사람이 다시 참고할 수 있는 맥락으로 바꾸는 것을 제 경쟁력으로 삼고 있습니다. AX 시대에는 많이 아는 것보다 경험, 정보, 실험을 재사용 가능한 구조로 정리해 다음 문제 해결에 연결하는 능력이 중요하다고 봅니다. 이 README는 그 방식을 보여주는 포트폴리오 진입점입니다. edu 아카이브, GitHub 레포, Hugging Face 모델/데이터셋/Spaces, W&B 공개 프로젝트를 연결해 전공 학습, 캡스톤/해커톤/부트캠프, 서비스형 프로젝트, 모델 배포와 실험 추적이 어떻게 검토 가능한 증거와 도구로 축적됐는지 보여줍니다. 저는 문제를 끝까지 이해하려는 태도를 기록과 구현으로 남기며, AI와 함께 활용 가능한 지식 시스템을 만드는 개발자로 성장하고 있습니다.
| Area | What to review first | What it demonstrates |
|---|---|---|
| AI / ML / LLM | large-language-models, LG Aimers archive, lg-aimers-8th-llm-compression, Hugging Face | AI 모델을 정확도뿐 아니라 평가 환경, 추론 효율, 경량화, 서빙 가능성 기준으로 이해합니다. |
| Data / Backend | big-data-analysis, Bigdata_Proj, cafeProj, GovOn datasets | SQL, 데이터 처리, Spark/Delta/MLflow, FastAPI, 서비스형 백엔드 흐름을 연결합니다. |
| Platform / Infra | operating-systems, container-orchestration, UMMAYA, GovOn Runtime | OS, Linux, Docker/Kubernetes, CUDA, Homebrew/npm 배포, HF Space 런타임 운영을 다룹니다. |
| Software Engineering | aioss-open-source-delivery, sw_2nd, homebrew-ummaya, algorithm workspace | GitHub Flow, CI, JSON-RPC 전환, 패키징, 알고리즘 문제 관리까지 구현 단위로 남깁니다. |
| Communication / Values | creative-writing, 창의적 글쓰기 제출물, 커뮤니케이션 가치관 근거, 고전읽기 제출물, 책과 미디어 제출물 | 기술을 효율성만이 아니라 자유, 공론장, 타인의 아픔, 삶의 의미, 책임 있는 전달의 관점으로 해석합니다. |
| Track | Evidence | What it proves |
|---|---|---|
| Product / Open Source | UMMAYA, homebrew-ummaya, UMMAYA docs, UMMAYA vision | Agentic RAG harness 구조를 조사·분석해 UMMAYA에 적용했습니다. DX 중심의 국가 인프라를 AX 목적의 query surface와 tool gate 구조로 마이그레이션하고 있습니다. |
| UMMAYA Research / Presentation | KSC 2026 presentation, research plan, KSC paper outline, UI wireframes | KSC 2026 발표·논문 제출 목표, 공공AX 정책 정합성, 권한 게이트, TUI 발표 화면을 문서와 실행 가능한 와이어프레임으로 정리했습니다. |
| Bootcamp / Hackathon | LG Aimers archive, lg-aimers-8th-llm-compression, MLSummerBootcamp | LG Aimers를 부트캠프와 해커톤 경험으로 연결하고, EXAONE-4.0-1.2B 경량화, vLLM 평가 환경, ML 기초 실습을 남겼습니다. |
| Capstone / Collaboration | DigitalPublishing, degree portfolio, GovOn presentation | 글로컬연합대학 Field 연합전공, 실증적 AI 개발 프로젝트(종합설계), GovOn 팀 발표를 협업 경험으로 연결합니다. |
| Model / Dataset / Runtime | HF models, HF datasets, HF Spaces | GovOn/EXAONE LoRA, AWQ, merged model, 한국어 민원/법률 instruction-tuning 데이터셋, Docker 기반 Spaces 런타임을 관리합니다. |
| Data / Dashboard Project | Bigdata_Proj | YouTube Data API v3 기반 K-POP 인기도 분석을 Spark, Delta Lake, MLflow, Streamlit 파이프라인으로 구성했습니다. |
| Applied / Service Projects | cafeProj, sw_2nd, LLM_Project | Next.js/FastAPI 카페 서비스, REST에서 JSON-RPC로 옮기는 소프트웨어 훈련, RAG/Text2SQL/LLM API 실습을 정리했습니다. |
| AI / Generative Media | lora_Proj, comfyui-models | OneTrainer, ComfyUI, SDXL/LoRA, 이미지·비디오 워크플로우를 서버 훈련과 모델 자산 관리로 연결했습니다. |
| Core CS Implementations | Artificial_Intelligence, ML, OS, cudaProj, Graphics, COTEST, 과제 제출물 근거 인덱스 | AI, ML, 운영체제, CUDA, 그래픽스, 코딩테스트 구현과 과제 증적을 같은 계열의 구현/학습 흔적으로 정렬해 검토 가능합니다. |
| Experiment Tracking | W&B public projects | GovOn, EXAONE 민원 분류, QLoRA 탐색, 재학습, 평가 프로젝트 9개를 공개 실험 추적 표면으로 연결합니다. |
| Path | Purpose |
|---|---|
| ComputerScience/00_graph-interfaces | Knowledge graph interfaces, stage maps, bridges, and course-level indexes |
| ComputerScience/01_programming-foundations | Coding basics, Python, Java, data structures, coding test prep |
| ComputerScience/02_math-theory | Probability/statistics, discrete math, optimization math, mathematical logic |
| ComputerScience/03_ai-ml-data | AI, ML, neural networks, computer vision, LLM, big data analysis |
| ComputerScience/04_systems-infrastructure | Linux, computer architecture, OS, networks, CUDA, Docker/Kubernetes |
| ComputerScience/05_software-engineering | Web, database, OSS, programming languages, AIOSS delivery |
| ComputerScience/06_algorithms-graphics | Algorithm design/analysis and computer graphics |
| ComputerScience/07_professional-humanities | Creative writing, intellectual property, classics, degree portfolio |
| LGAimer | LG Aimers certificate, LLM compression materials, EXAONE/vLLM notes, application drafts |
| certifications | Information processing and data-analysis certification prep records |
Last updated: 2026-05-29