Research Scientist - Environments, Data and Post-Training
Core
Develop novel post-training methods, training recipes, and evaluation systems to improve frontier AI model reasoning, tool use, and agentic behavior.
Role type
Senior IC research scientist (post-training & RL)
Builds
Scalable post-training pipelines, RLVR systems, benchmarks, and evaluation rubrics for frontier models
Domain
Artificial Intelligence / Machine Learning / Reinforcement Learning
Deliverable
production ML models
Required skills
Post-training methods, Reinforcement Learning, Language-model evaluation, Data-centric ML, Experimental design, Python programming, Distributed systems
Preferred skills
Industry post-training experience, Top-tier conference publications, Synthetic-data generation, Large-scale evaluation infrastructure, Cloud infrastructure
Technologies
GRPO, DAPO, RLVR, Open-source tools
Responsibilities
Implement novel post-training methods, Design and run experiments across datasets and reward functions, Build scalable data generation and filtering pipelines, Develop rubrics and scoring systems for training decisions, Investigate model capabilities and failure modes, Contribute to open-source post-training tools
Seniority
Senior, hands-on IC
