- Remote, United States
- Remote
- Full-time · Senior
- 8+ years
Design, optimize, and scale infrastructure for GenAI and LLM workloads, focusing on GPU profiling, distributed training, and high-performance compute environments.
Responsibilities
- Design and implement scalable infrastructure for LLM and GenAI workloads across multi-GPU environments
- Perform GPU profiling, benchmarking, and performance optimization for distributed training workloads
- Manage and schedule compute-intensive jobs using Slurm-based clusters and OpenShift/Kubernetes environments
- Enable and optimize the NVIDIA GPU stack (CUDA, cuDNN, NCCL, Triton, RAPIDS, etc.)
- Collaborate with cross-functional teams to deploy models in research and production environments
Requirements
- Nice to have: NVIDIA NIMs
- Nice to have: DGX systems
- Nice to have: LLMOps frameworks
- Nice to have: MLOps integration
- Nice to have: Vector databases