Member of Technical Staff - Research

What you'll do

• Turn ML research problems into environments that reward meaningful progress and sound research judgment.
• Test whether environment rewards are repeatable and track the quality of scored work; build systems to create and validate environments at scale.
• Run post-training experiments to identify trainable environments.

Nice to have

• Experience with LLMs, RL, RLHF/RLAIF, RLVR, post-training, evaluations, graders, synthetic data, model training, or coding agents at a frontier/data lab.
• Understanding trustworthy reward evaluation and model reward exploitation; experience designing automated environment-building and quality-checking systems.

Tools

ML, LLMs, RL, RLHF/RLAIF, RLVR

AI focus

Build research environments and evaluations to improve agent capabilities and test whether frontier models can ideate, experiment, and produce impactful research.