RL Post-Training Systems Engineer (Distributed Infra)

NVIDIA; New York (NY); On-site; USD 184,000 - 357,000; Full time

Job summary

NVIDIA is building an RL post-training infrastructure team to scale experimental workflows from a single GPU to thousands of nodes, delivering reliable, high-performance runtimes for researchers. You will collaborate with researchers and labs, optimize PyTorch-based RL loops, and improve fault tolerance, elastic scaling, and portability across CPU, GPU, and LPUs.

This role offers the chance to contribute to VeRL, Miles, TorchTitan and related ecosystems while shaping production-grade distributed

Qualifications

• MS or PhD in CS/CE or related field.
• 5+ years of distributed systems or ML infrastructure experience.
• Strong Python and C/C++ proficiency.
• Experience building/operating large-scale distributed systems in production.
• Excellent verbal and written communication across teams.

Responsibilities

• Architect and build RL post-training infrastructure at scale from a single GPU to thousands of nodes.
• Tune RL training–inference–rollout loops for performance on GPUs, CPUs and LPUs.
• Improve open-source RL frameworks and collaborate with hardware and software teams.
• Focus on fault tolerance, elastic scaling, and fast restarts for long-running jobs.

Skills

Python

C/C++

Distributed systems

Communication skills

Education

MS or PhD in CS/CE or related field

Tools

PyTorch

Ray

Kubernetes

Job description

NVIDIA is building an RL post-training infrastructure team to scale experimental workflows from a single GPU to thousands of nodes, delivering reliable, high-performance runtimes for researchers. You will collaborate with researchers and labs, optimize PyTorch-based RL loops, and improve fault tolerance, elastic scaling, and portability across CPU, GPU, and LPUs.

This role offers the chance to contribute to VeRL, Miles, TorchTitan and related ecosystems while shaping production-grade distributed