CV

Contact Information

Name MinJu Jeon
Professional Title AI Researcher
Email mnju5026@naver.com
Phone +82 010 8975 5026
Location Seoul,

Professional Summary

AI researcher with hands-on experience in large-scale multimodal model training and data-centric optimization. My work spans vision-language pre-training, post-training refinement (data curation, synthetic caption generation, retrieval-augmented learning), and production deployment of multimodal systems.

Experience

  • 2025 - present

    Seongnam, South Korea

    Research Intern, Voice Tech
    NAVER Cloud
    • Developed robust multilingual G2P model for noisy, non-standard inputs
    • Built G2P benchmark for non-canonical words across KO, EN, and VI
    • Analyzed user corpora to identify failure modes in standard G2P systems
  • 2024 - present

    Seoul, South Korea

    Research Assistant
    Hanyang Multi-Model AI Laboratory
    • Researching Saliency-Aware Video Reweighting and Retrieval-Augmented Learning for DVC
    • Collaborated with Piaspace on industrial research projects for automated highlight extraction
  • 2023 - 2024

    Seoul, South Korea

    Undergraduate Research Intern
    Hanyang Multi-Model AI Laboratory

Education

  • 2024 - 2026

    Seoul, South Korea

    M.S.
    Hanyang University
    Data Science
    • Advisor: Dong-Jin Kim
  • 2020 - 2024

    Seoul, South Korea

    B.S.
    Hanyang University
    Industrial Engineering, Big Data Science (Double Major)

Publications

  •
    Sali4Vid: Saliency-Aware Video Reweighting and Adaptive Caption Retrieval for Dense Video Captioning
  •
    Cap4Bridge: Caption-Guided Cross-Modal Contextualization with Stochastic Augmentation for Text-Video Retrieval
  •
    Enhancing Lightweight Image Captioning with Localized Features and Keyword Extraction
  •
    Time-aware Video Frame Reweighting and Captioning Tool
  •
    Follow the Saliency: Supervised Saliency for Retrieval-augmented Dense Video Captioning
  •
    SAIL: Similarity-Aware Guidance and Inter-Caption Augmentation-based Learning for Weakly-Supervised DVC
  •
    SynC: Synthetic Image Caption Dataset Refinement with One-to-many Mapping for Zero-shot Image Captioning

Projects

  • Retrieval-Augmented Dense Video Captioning and QA System
    • Funded by Piaspace
    • Built an end-to-end pipeline for automated highlight extraction and dense captioning of long-form videos (1hr+)
    • Developed a retrieval-augmented QA module enabling natural language queries over video content
  • Risk State Prediction Model Using Construction Site Images
    • Funded by Doosan Enerbility
    • Developed a VLM-based risk-state prediction system fine-tuned to detect unsafe worker behavior and equipment hazards
  • Zero-Shot Captioning for Driver Status Reporting Agents
    • Funded by Hyundai NGV
    • Proposed a zero-shot image captioning method for in-vehicle driver monitoring to generate natural language descriptions of driver states

Skills

:
:
:
:

Awards

  • 2025
    Best Paper Award (Bronze)
    IEIE Summer Annual Conference
  • 2024
    Outstanding Poster Presentation Award
    IEIE Summer Annual Conference
  • 2022
    Bronze Prize
    DACON AI Competition (Mandarin Orange Yield Prediction)

References

  • Dong-Jin Kim