Founding Research Scientist

Design and train reinforcement learning systems using RLHF, RLAIF, and reward modeling approaches for scientific hypothesis generation and evaluation. Develop process reward models and verifiers to provide fine-grained supervision over intermediate reasoning steps. Contribute to alignment and oversight research for complex scientific tasks where ground truth is expensive or ambiguous. Build robust training pipelines, run large-scale experiments, and iterate quickly across the research-to-production lifecycle. Contribute to meaningful benchmarks and evaluation methods for domain-specific reasoning. Own problems end-to-end from idea through experimentation to production.