Senior Software Engineer, RL Post-Training Frameworks
Architect and develop scalable reinforcement learning post-training infrastructure that operates across GPUs, CPUs, and LPUs, from single-node experimentation to multi-thousand-node production systems. Work closely with AI researchers and infrastructure teams to optimize distributed training-inference-rollout loops, improve open-source RL frameworks, and integrate with distributed runtimes like Ray and Monarch. Focus on performance, fault tolerance, elastic scaling, and hardware-aware execution for next-generation AI workloads.