Research Engineer – Reinforcement Learning (RL) Systems & Infrastructure (Seed Infra)

Build end-to-end reinforcement learning systems for large-scale models, including rollout, training, evaluation, and deployment pipelines. Develop scalable, fault-tolerant RL infrastructure that handles dynamic workloads and heterogeneous compute. Optimize distributed training across GPU clusters to improve throughput, utilization, and stability. Collaborate with researchers on system–algorithm co-design and create tooling for monitoring, debugging, and observability of large-scale RL training.