Spellbrush
AI Anime Researcher - LLM
San Francisco or Tokyo · Full-time · Posted Sep 2, 2025
- Visa sponsorship
- Visa sponsorships are available for qualified candidates.
About the job
Research and develop LLM systems for a 3D adventure game where an AI companion participates in gameplay and narrative.
What you'll do
- Develop and evaluate LLM post-training approaches for the game’s AI companion.
- Gather and use human feedback, including feedback from interaction logs.
- Build LLM infrastructure and balance research exploration with product-focused development.
What they ask for
- Published work or industry experience with LLM post-training.
- Deep understanding of LLM training and evaluation.
- Experience gathering human feedback.
- Ability to balance research exploration with product-focused development.
Nice to have
- Post-training focused on roleplay; latency optimization experience.
- Experience with LLM infrastructure, such as data processing backends, checkpointable data loaders, sharding, memory-saving techniques, scalable benchmarking, or checkpoint selection.
Tools
LLM, RLHF
AI focus
Post-train, evaluate, and optimize LLMs for a real-time, gameplay-integrated AI companion; apply human feedback and build supporting training infrastructure.
Model-written summary · the original posting is authoritative