Summer 2027 Master's AI Research, Reinforcement Learning and LLM Post-Training Intern

Conduct reinforcement learning research for post-training language and code models, including policy optimization, preference learning, reward modeling, and exploration/credit-assignment methods. Build and run controlled, verifiable experiments using preference-based or simulator feedback, diagnose failure modes like reward hacking and policy degeneration, and develop evaluation methods aligned with engineering constraints. Partner with research and infrastructure teams on rollout generation, training, logging, and reproducibility while documenting results in technical reports and publications.