Staff Engineer, ML Environments & Post-Training

Hillclimb is a research-focused startup aiming to improve agent capabilities toward recursive self-improvement. Our team designs environments and evals that push frontier models to ideate, experiment, and produce meaningful research progress. We value expertise in LLMs, RL, RLHF/RLAIF, and automated evaluation systems to ensure trustworthy rewards. This role emphasizes building scalable pipelines and validating environments before training, with a collaborative frontier/data lab setting.