Yuji Naraki
Research Engineer in Tokyo, Japan
LegalOn Technologies, Inc.
I'm a Machine Learning Engineer and Researcher based in Tokyo, Japan, with a focus on model editing and natural language processing. I'm passionate about bridging cutting-edge research with real-world applications, and I take pride in owning projects from start to finish. I'm also highly interested in technologies related to Personalized Companion Agents — interactive systems that engage naturally with humans. I thrive in collaborative environments and enjoy continuously learning new technologies and tools to solve complex problems.
Publications
International Conference on Learning Representations (ICLR) 2025 2025
Mastering Task Arithmetic: τJp as a Key Indicator for Weight Disentanglement
Kotaro Yoshida, Yuji Naraki, Takafumi Horie, Ryosuke Yamaki, Ryotaro Shimizu, Yuki Saito, Julian McAuley, Hiroki Naganuma
NeurIPS 2024 Workshop: Fine-Tuning in Modern Machine Learning: Principles and Scalability (FITML) 2024
Mastering Task Arithmetic: τJp as a Key Indicator for Weight Disentanglement
Kotaro Yoshida, Yuji Naraki, Takafumi Horie, Ryosuke Yamaki, Ryotaro Shimizu, Yuki Saito, Julian McAuley, Hiroki Naganuma
The 13th Conference on Language Resources and Evaluation 2022
Evaluating the Effects of Embedding with Speaker Identity Information in Dialogue Summarization
Yuji NARAKI, Tetsuya SAKAI, Yoshihiko HAYASHI
Preprints
preprint 2024
Augmenting NER datasets with LLMs: towards automated and refined annotation
Yuji Naraki, Ryosuke Yamaki, Yoshikazu Ikeda, Takafumi Horie, Kotaro Yoshida, Ryotaro Shimizu, Hiroki Naganuma
Experience
ML Research Engineer — LegalOn Technologies, Inc. (formerly LegalForce, Inc.)
Developed stateless APIs using ML models for legal document analysis. Developed asynchronous APIs using Cloud PubSub for time-consuming LLM processing.
Research Internship — LegalForce, Inc.
Researched on machine translation for texts with technical terms.
Research Engineering Internship — CyberAgent, Inc.
Developed a text-generation engine for advertisement using templates.
Data Scientist Internship — PELORIA, P.B.C.
Developed a deep learning model that ran on social networking sites to create a city risk index.
Machine Learning Engineer Internship — AlgoAge Inc
Developed iOS and web applications for attendance & health management. Developed a web application for image annotation. Developed algorithms and frameworks for face recognition. Improved generation of animated characters with face mesh input. Implemented extraction of images from a 3D camera.
Research Assistant — Science Tokyo (formerly Tokyo Institute of Technology)
Surveyed NeurIPS 2019 papers in the field of deep learning for language. Implemented an environment for large-scale experimentation in language model learning.
Machine Learning Engineer Internship — Logbii Inc
Analyzed videos of object detection for labeling. Developed a SaaS application for machine learning. Analyzed dynamic dashboard camera for collision detection.
Education
Waseda University
Master of Engineering in Computer Science
Advisor: Prof. Yoshihiko Hayashi and Prof. Tetsunori Kobayashi
Thesis: End-to-End Length Controllable Summarization using Soft Information Masking（Original Title: 指定要約長に応じたソフトな内容選択による要約長操作可能なEnd-to-End要約）
Waseda University
Bachelor of Engineering in Computer Science
Advisor: Prof. Tetsuya Sakai
Thesis: Evaluating the Effects of Embedding with Speaker Identity Information in Dialogue Summarization