Senior DGX Cloud AI Infrastructure Software Engineer
Core
Designing, building, and maintaining AI infrastructure software to optimize efficiency and resiliency for large-scale AI pre-training, post-training, and inference workloads.
Role type
Senior IC AI infrastructure software engineer
Builds
Infrastructure software and tools for NVIDIA's DGX Cloud AI platforms
Domain
High-performance computing and AI infrastructure
Deliverable
production ML models | infrastructure
Required skills
Large-scale distributed systems development, observability platform integration, root cause analysis across application to hardware levels, software engineering practices (testing, CI/CD), Python/C/C++ programming
Preferred skills
RDMA software stack (NCCL, IB verbs, ucx), deep learning framework internals (PyTorch, TensorFlow, JAX, Ray), datacenter-scale failure analysis
Technologies
Python, C/C++, ELK, Prometheus, Loki, NCCL, IB verbs, ucx, libfabrics, PyTorch, TensorFlow, JAX, Ray
Responsibilities
Develop infrastructure software and tools for large-scale AI training and inference, optimize tools and libraries for efficiency and resiliency, co-design and implement APIs for resiliency stacks, define reliability metrics to track system performance, analyze and triage failures from application to hardware levels
Seniority
Senior, hands-on IC
