Senior Software Engineer, RL Post-Training Frameworks

Architect and build reinforcement learning (RL) post-training infrastructure that scales from single-GPU experimentation to production across thousands of nodes. Tune training–inference–rollout loops across GPUs, CPUs, and LPUs; improve performance and usability of open-source RL frameworks; and deliver fault tolerance via elastic scaling and fast restarts. Partner with researchers and framework owners while collaborating with NVIDIA networking, math, compiler, and hardware teams to meet researcher needs.