Reinforcement Learning Research Engineer

Google DeepMind is hiring a Senior RL Research Engineer to work on reinforcement learning for language model alignment and reasoning. You will develop new RL algorithms that improve the capabilities and safety of our foundation models. This role involves designing reward models, implementing RL training loops at scale, and developing new techniques for improving model reasoning through self-play and multi-agent interactions. The ideal candidate has deep expertise in RL theory and practice, combined with strong engineering skills for implementing algorithms at scale.

- 5+ years in RL research or engineering - Strong background in RL theory and algorithms - Experience with RLHF and language model alignment - Expert JAX or PyTorch skills - Published research in RL or related areas - PhD in ML/RL preferred