Junya Kataoka

LLM Community Researcher, LLMATCH

Selected for the third and fourth cohorts of LLMATCH, a student research support program hosted by the LLM Community at the Matsuo-Iwasawa Lab, The University of Tokyo. Since October 2025, I have been conducting research on LLMs that imitate specific individuals and characters.

Intern at EAGLYS Inc.

Implemented a TF-IDF-based retrieval function for an internal system at an insurance company to search for similar past inquiries.

Graduate School of Advanced Science and Engineering, Waseda University

Department of Electrical Engineering and Bioscience / Bannai Laboratory

School of Advanced Science and Engineering, Waseda University

Department of Electrical Engineering and Bioscience / Bannai Laboratory

Automated Tracking System for Lipid-Labeled Quantum Dots in Microscopy Videos

Developed an automated tracking system for lipid-labeled quantum dots in microscopy videos by constructing a synthetic data generation pipeline from a few annotated images, training YOLOv12 for detection, and improving ByteTrack for more consistent tracking of small and irregularly moving particles.

Progate Hackathon powered by AWS

Worked on the implementation of an image RAG pipeline using Amazon Bedrock Knowledge Bases.

Kaggle Competition: Deep Past Challenge

Ranked 665th out of 2,674 teams (top 25%) in the Deep Past Challenge, a machine translation competition focused on translating Old Assyrian Akkadian cuneiform texts into English.

Fundamental Information Technology Engineer Examination (基本情報技術者試験)

TOEIC 860

Deep Generative Models 2026 Spring

Large Language Model Course (Advanced) 2025 Autumn

Large Language Model Course (Basic) 2025 Autumn

GCI 2025 Summer