Active and Passionate MLOps Engineer

Github: https://github.com/7dudtj

e-mail:

Call:

I was attracted to creating business value through personalized recommendation, so I started studying the recommendation system, and I completed the boostcamp AI Tech 4th of the Naver Connect Foundation for deeper learning.

Projects such as <HiRM>, <My Own House - Personalized Interior Product Recommendation>, <Movie Recommendation>, <Deep Knowledge Tracing>, <Book Rating Prediction>, and <Stock Price Prediction> have been carried out. For more information, please check the link in Project.

Through activities in various clubs and organizations, I am trying to contribute to the college student algorithm community.

Education

• Yonsei University major in Computer Science (2018.03 ~ 2024.02)

Career

• IT Engineer at NH Investment & Securities (2023.12 ~ )

• Intern at SNECompany Corporate Research Institute (2022.06 ~ 2022.08)

  • Start-up of ugly agricultural product direct trading platform using AI technology
  • I have established an environment for performing vulnerability analysis to enhance the security of in-house programs
  • Spotbugs and PMD extension were applied in the Eclipse environment
  • I identified security vulnerabilities in the in-house program and presented necessary measures

Project

• HiRM (2023.03.30 ~ 2023.05.31)

  • High Speed Recommendation Model based on Graph Signal Processing
  • A high-speed recommendation model designing project based on graph signal processing
  • Based on the GF-CF paper learning, I was in charge of graph signal processing and introduced Diversity metric
  • Compared to the GF-CF model on which HiRM was based, Recall@20 improved by 2-3% and NDCG@20 by 3-4%, and Diversity significantly improved

• My Own House - Personalized Interior Product Recommendation (2023.01.09 ~ 2023.02.03)

  • It is a project to recommend personalized interior products based on the interior style and products that users prefer
  • I led the project as PM and designed the system architecture of the project
  • I was responsible for database deployment (MySQL), project deployment and operation (GCP), and project version management (Git)
  • I have taken measures against various security vulnerabilities for the security of the project
  • In order to maximize the development capabilities of team members, I established, documented, and shared various conventions necessary for the project

• Movie Recommendation (2022.12.12 ~ 2023.01.06)

  • The performance analysis of the model was conducted quickly using the RecBole library, and the prediction performance was improved by directly devising the ensemble algorithm (Recall@10 0.1603 → 0.1626)