About Junichiro

As a Solutions Architect at NVIDIA, I work with Japanese enterprise customers on large language model training and inference, with a focus on the NeMo framework. I joined NVIDIA in February 2026.

Previously, as a Researcher at Preferred Networks Inc., I primarily engaged with clients in the healthcare and life sciences sectors and served as the tech lead for providing BtoB solutions using Large Language Models (LLMs).

I led a research project to build a medical domain-specialized LLM, resulting in Llama3-Preferred-MedSwallow-70B and Preferred-MedLLM-Qwen-72B, the first open LLMs that surpass GPT-4 in the Japanese Medical Licensing Exam (JMLE). I also led a project for developing a deep learning based solution aimed at enhancing the diagnosis of endometriosis through the analysis of MRIs, in collaboration with a leading pharmaceutical company.

I have experience formulating problems through consultations with clients and providing solutions based on deep learning / machine learning. I have also been the main mentor for four research interns and one part-time engineer.

During my PhD, I worked on the development of machine learning methods for gene expression/mutation data to tackle the problem of antibiotic resistance of bacteria.

CV

Junichiro Iwasawa (岩澤 諄一郎)