about 6 hours ago
Tokyo, JapanMid Level / Senior
H1B Sponsor
Responsibilities
- Design and execute post-training strategies for language and multimodal models.
- Build and curate high-quality training data with a focus on Japanese-language capabilities.
- Develop evaluations to identify capability and reliability gaps.
- Conduct systematic error analysis to improve training methods and model behavior.
- Run controlled experiments and communicate clear recommendations.
- Develop reliable training and evaluation pipelines in collaboration with infrastructure teams.
- Contribute methods and findings to accelerate post-training work.
Requirements
- Hands-on experience with post-training modern language or multimodal models.
- Strong understanding of machine learning fundamentals and current post-training methods.
- Solid engineering skills and proficiency with the open-source ML ecosystem.
- Experience designing and running rigorous experiments, including systematic error analysis.
- Ability to turn research ideas into reliable implementations and measurable improvements.
- Proficiency in English for collaboration with global teams.
- Experience leveraging agents to enhance work.
Benefits
- High-impact model work improving foundation models in real products.
- Direct access to core teams for close collaboration.
- Ownership in a growing market with strong customer demand.
- Supportive culture prioritizing continuous learning and mentorship.
- Primarily remote work with flexible hours and unlimited paid time off.
- Competitive salary, equity in a unicorn-stage company, and standard benefits.