Junyeop Na
Senior Machine Learning Engineer
ML Engineer with 4+ years of experience building various AI solutions. Specialized in using deep learning, computer vision, and building LLM agent/RAG systems.
E X P E R I E N C E
ML Engineer
NAVER Media AI
AUG 2025 - PRESENT
TensorFlow
PyTorch
TensorRT
Triton- Developed BestCrop, a computer vision system that identifies the best region of an image for thumbnail cropping.
- Developed BestCut, combining CLIP-style multimodal embeddings with aesthetic scoring to select visually appealing thumbnails that match a post's content.
- Developed AI models for SPORTER that provide contextual signals for automated clip and highlight generation across the 2026 World Cup, MSI, EWC, Asian Games, and LoL Worlds.
F E A T U R E D P R O J E C T S
SPORTER - AI Highlight Generation System for Soccer and Esports
Developed AI models for the SPORTER System: automated realtime clips/highlight generation system for CHZZK's 2026 World Cup, MSI, EWC, Asian Games, collecting 100 million total views.
In the news
Public reporting on the system. Figures reflect the dates cited below.
- Aju PressAsian Games AI clips reach 53.2 million views
Reports more than 4,600 AI clips and 53.2 million views as of October 4, with SPORTER supporting highlight creation across the Asian Games.
(opens in a new tab) - Business Watch (Korean)SPORTER brings AI highlights to the Asian Games
Covers SPORTER's use on CHZZK, including AI clips and chapters, alongside multi-source highlights that group footage by athlete, team or topic.
(opens in a new tab) - CEOSCOREDAILY (Korean)NAVER develops SPORTER ahead of the World Cup
Early reporting describes development of an AI system to detect key moments during matches and generate highlight clips in real time.
(opens in a new tab)
