Member of Technical Staff - ML Scientist, Japanese Multimodal

Highlights

• Improve Japanese multimodal foundation models used in real deployments.
• Own full post-training loop from data to high-quality checkpoints.
• Turn deployment signals into broader model quality improvements.

What they offer

• Competitive salary
• Equity in unicorn-stage company
• Standard benefits for employees in Japan
• Unlimited paid time off
• Flexible working hours

Ideal candidate

Experimental ML scientist who loves turning data insights into working models. You thrive owning the full loop from hypothesis to checkpoint and collaborating across global research teams.

Requirements

• Hands-on experience post-training modern language or multimodal models
• Strong understanding of ML fundamentals and current post-training and RL methods
• Solid engineering skills and proficiency with open-source ML ecosystem
• Experience designing rigorous experiments with baselines, ablations and error analysis
• Experience building, curating or assessing training and evaluation data at scale
• Ability to turn research ideas into reliable implementations and measurable improvements
• Proficiency in English for collaboration with global teams
• Experience leveraging agents to amplify your own work
• Experience with preference optimization or RL methods for foundation models
• Experience post-training multimodal models involving text, vision, or audio