Jaebok Lee

Experience

• 

  Nanyang Technological University Singapore

  Aug 2018 - Dec 2018

  Research Assistant

  • Korean Sentiment Resource KSenticNet for Sentiment Analysis (2018) - https://github.com/zzaebok/ksenticnet - Built Korean sentiment analysis resource, KSenticNet, by combining SenticNet framework and KAIST Korean WordNet, and released it as open-source

• 

  Samsung Electronics

  Jun 2019 - Aug 2019

  Research Engineer Intern

  • Battery SoC estimation (2019) - Developed battery State-of-Charge estimation model using LSTM, incorporating Monte Carlo Dropout for safety and reliability

• 

  Samsung Electronics

  Feb 2020 - now

  Machine Learning Engineer

  • Neural Machine Translation Model Development, Language Intelligence Team (2024) - Developed Galaxy on-device machine translation model with binary quantization, enhancing performance through back translation and knowledge distillation - Fine-tuned Samsung Gauss (LLM) for multilingual machine translation with preference optimization and LoRA• Knowledge Base Question Answering System, Data Analytics Lab (2023) - Developed natural language to SPARQL query translation model using constrained decoding - Collaboration with Samsung Mobile Experience division• Short-text Application Development, Data Research Team (2022) - Developed and deployed short-form text reading mobile application `Book-ing' aimed at improving adolescent literacy - Frontend development (Flutter) - Backend development (FastAPI, kubernetes, MongoDB, redis) - Collaboration with Samsung R&D Institute India-Delhi• Time-adaptive App Usage Prediction, Data Research Team (2021) - Designed time-adaptive app usage prediction algorithm by applying Graph Neural Networks on a user-specific app usage transition graph, enabling post-deployment adaptability - Collaboration with Samsung Electronics Mobile Division• Knowledge Graph based Explainable Food Recommendation Service, AI Service Lab (2020) - Developed knowledge graph-based explainable food recommendation service, an internal company tool that recommends cafeteria menus and explains reasoning behind user preferences - Backend API development (kubernetes, gRPC) Show less