Biography

MS in Data Science, Seoul National University , Seoul, Korea (Image generation/translation, Video understanding, Multimodal AI, LLM)
BSc in Computer Science & Mathematics, Korea University , Seoul, Korea

Oct 2022 – Present: AI Engineer

• Virtual try-on, generative image pipelines, diffusion model customization
• Digital asset generation for various industries
• Production-ready AI model deployment and workflow optimization
• Large Language Model (LLM) finetuning & deployment (domain-specific instruction tuning, RAG system optimization, inference acceleration)
• Small Language Models (SLM) optimization (parameter-efficient tuning, quantization)

Jul 2022 – Oct 2022: Opensource Engineer (Intern), Quansight Labs

• Contributed to PyTorch-Ignite open-source project
• Improved metric testing framework and optimized distributed metric calculation
• Addressed CI issues and resolved community-raised bug reports

Dec 2020 – Jul 2021: Machine Learning Engineer, 11Street (E-commerce)

• Improved information retrieval results (query clustering, automated curation)
• Optimized search models and enhanced large-scale user query performance

Nov 2019 – Dec 2020: AI Engineer, Carrot Insurance

• Developed deep learning-based crack detection model for mobile insurance
• Built dog breed classification model for pet insurance products

Jul 2016 – Oct 2019: Data Analyst, Samsung Electronics (Mobile Division AI Team)

• Performed statistical analysis on user data (Samsung Pay, Bixby)
• Increased monthly active users by 15% through service optimization and market analysis

Jul 2013 – Jun 2016: Procurement Manager, Samsung Electronics (Procurement Team)

• Semiconductor demand forecasting (KNN clustering, regression modeling)
• Achieved >10% stock reduction and improved Just-In-Time performance

Research Interest

• Generative modeling (diffusion models, virtual try-on)
• Video understanding
• Multimodal AI
• Large Language Models (LLM)

Patents & Intellectual Property

• KR 10-2023-0128093 (2023) “Image Generation Apparatus and Method”
• KR 10-2023-0120201 (2023) “User Terminal, Server & Method for Image Editing”
• KR 10-2023-0116928 (2023) “Image Editing Apparatus and Method”
• US PFP230066US (2023) “User Terminal, Server & Method of Operation for Image Editing”
• KR 10-2022-0116656 (2022) “Virtual Try‑On Method Supporting Diverse Poses & Body Types”
• KR 2202-068959 (2022) “Virtual Try-on Program Development”

Publications

Peer-Reviewed

• Lee et al. (First author), Learning to Wear: Details-Preserved Virtual Try-on via Disentangling Clothes and Wearer, BMVC 2022
• Lee et al. (First author), Towards Detailed Characteristic-Preserving Virtual Try-On, CVPR Workshop 2022
• Chae et al. (Co-author), Towards a Complete Benchmark on Video Moment Localization, AISTATS 2024
• Mun et al. (Co-author), BaSSL: Boundary-aware Self-Supervised Learning for Video Scene Segmentation, ACCV 2022

Highlights

• Advanced virtual try-on and image generation pipeline development
• Proficient in Python, PyTorch, TensorFlow, Keras, and Scikit-Learn (Intermediate C++ on Linux/Windows)
• Strong mathematical background (Statistics, Linear algebra, Numerical analysis, Differential equations)
• Experience handling structured and unstructured data (image, video, text)
• Implementation of state-of-the-art ML/DL papers (github.com/puhuk)
• OpenUP (Korea National Open Source Support Center) Frontier Developer (Jun. 2021 ~ )
• AI bootcamp deep learning project mentor (Ministry of Employment and Labor w/ Samsung Multicampus & SK) (Nov. 2019 ~ )
• TensorFlow Developer Certificate (Oct. 2020)
• Sales forecast competition top 10% model (Samsung Electronics) (Sep. 2019)