|
Junjie Chen I'm an applied scientist working on multimodal LLM systems and LLM evaluation. I enjoy learning from the product development cycle: understanding the requirement, building the system, and evaluating and triaging the true requirement. I'm currently a postdoctoral researcher at the University of Oxford, where I build and evaluate end-to-end comic translation. In parallel, I am developing an audio-only heart-rate monitoring algorithm for Marley Health. Before that I was an Amazon Science Fellow working on LLM manga translation, and a research intern at Rakuten working on explainable PII detection. I received my Ph.D. in Computer Science from the University of Tokyo. Email / Resume / Scholar / Semantic Scholar / Github |
|
Experience |
Postdoctoral Researcher, University of Oxford — May 2026 – present
Amazon Science Fellow, Amazon, Tokyo — Aug 2025 – Feb 2026 Research Intern, Rakuten, Tokyo — Nov 2024 – Feb 2025 Research Assistant, LLM-jp, National Institute of Informatics — Jun 2024 – Feb 2025 |
Education |
Ph.D. in Computer Science, The University of Tokyo — 2021 – 2025
M.S. in Computer Science, The University of Tokyo — 2019 – 2021 B.S. in Software Engineering, Northeastern University (China) — 2015 – 2019 |
Grants |
Travel Grant ($2,000), AAAI — 2025
Special Allowance for Outstanding Student, JSPS — 2024 DC2 Doctoral Fellowship, Japan Society for the Promotion of Science (JSPS) — 2023 IST-RA Fellowship, The University of Tokyo — 2022 |
My ResearchI am interested in Natural Language Processing, Multimodal Machine Translation, and Signal Processing. Currently, I want to understand the factor that makes human like or dislike a comic translation. My PhD research on unsupervised parsing is to find a measurable proxy to "syntax/semantic units in languages". Below are my 11 publications, including ICLR (Spotlight), ACL (Oral), NAACL, and AAAI (Oral). They're also on Google Scholar. |
|
Website template from Jon Barron. |