Hello! I’m an undergraduate student at Seoul National University, majoring in Economics and Computer Science. My research interests revolve around audio-visual multimodal learning and audio generation.

Education

• Seoul National University (Seoul, Korea)
  • BS in Economics, Computer Science
  • GPA: 3.98/4.3
  • Mar 2019 - Present

Work Experience

• Beautiful Noise Inc. (Seoul, Korea)
  • NFT Project Manager (Nov 2021 - Aug 2022)
  • Served as a pivotal team member within two international NFT projects(GENESIS,ESION), spearheading comprehensive project oversight and coordination.
• DeepTrade Technologies Corp. (Seoul, Korea)
  • Mobile App Developer (Jan 2023 - Mar 2023)
  • Engaged as a Full Stack App Developer proficient in Django and React Native, responsible for crafting API architecture and enhancing application functionality.
• SNU Computing and Memory Architecture Laboratory (Seoul, Korea)
  • Research Intern (July 2023 - November 2023)
  • Engaged in an in-depth investigation and research on the advanced field of neural radiance field.
• SNU Music & Audio Research Group (Suwon, Korea)
  • Research Intern (January 2024 - February 2024)
  • Implemented SingSong: Generating Musical Accompaniments from Singing paper by Google Research from scratch with modifications, utilizing Transformer models for music generation from vocal inputs.

Projects

• CT reconstruction
  • Implemented and assessed the feasibility of a CPU-based FDK algorithm, incorporated the characteristics of the half-fan detector into the weighting process, and improved image quality through the application of diverse filter functions.
  • Github Link, Project Link
• Improving Preprocessing for LipReading
  • Explored the effectiveness of a preprocessing technique aimed at enhancing lip segmentation to improve lipreading accuracy. Leveraged BlazeFace and BiseNet models to highlight lip regions in video data, subsequently training the LipNet model on this refined dataset.
• Open-SingSong :An Implementation of ‘SingSong: Generating Musical Accompaniments from Singing’
  • The paper titled ‘SingSong: Generating Musical Accompaniments from Singing’ by Google Research explores the creation of instrumental music to accompany input vocals using Transformer. Since the code for this paper wasn’t made open-source, I attempted to implement the described methods with a few modifications.
  • Github Link , Project Link

Awards

• Encouragement Prize, Thesis Contest (Korea Deposit Insurance Corporation, Oct 2021)
  • Recognized for proposing a novel method to detect operational risks by analyzing time series private equity prices.
• Excellence Prize, Fintech Idea Contest (Fintech Center Korea, Sep 2023)
  • Acknowledged for presenting a pioneering business concept that combines natural language processing and recommendation systems for an international stock trading platform.

English Skills

• New Teps (Feb 2023)
  • 534/600