English summary for screening — check the original posting before applying.
ELYZA, an AI startup from the University of Tokyo's Matsuo Lab, is seeking a Research Engineer to develop LLMs. You will leverage Japan's leading GPU resources and domestic model development to implement research results into societal applications, working on foundational models, VLM/VLA, AI agents, and more.
Must-haves
- 2+ years of Python development experience
- 1+ year of practical Machine Learning/Deep Learning experience
- Experience with team development using Git/GitHub
- Experience with container-based development (Docker, Kubernetes)
- Business-level Japanese communication skills
Nice-to-haves
- Experience with multi-node distributed learning (FSDP, DeepSpeed, Megatron-LM)
- Experience implementing Transformer architectures
- Experience implementing and evaluating LLM post-training (SFT/RLHF/DPO)
- Experience developing web applications or API servers
- Papers accepted/presented at ML conferences (NeurIPS, ICML, ICLR, ACL)
- Top rankings in data analysis competitions (Kaggle)
Tech stack
PythonShell ScriptAzureGoogle CloudAWSABCIDockerSingularityHugging FaceW&BNotionGitHubSlackGoogle MeetDiscord
Work style
Full remote within Japan is possible. Choice between remote/office work. Super flextime system (no core hours: 6:00-22:00).
Other notes
Annual salary: ¥6,000,000 - ¥9,800,000. Stock option system available. Side jobs permitted.