
Senior GPU Networking Architect
Posted May 20

Posted May 20
This is a fully remote position, open to applicants in Poland.
• Design, implement, and enhance GPU communication kernels that support collective and point-to-point operations in large-scale AI frameworks.
• Utilize extensive knowledge of GPU architecture—including thread scheduling, memory hierarchy, and execution pipelines—to enhance kernel performance, reduce latency, and synchronize computation with communication.
• Create GPU-resident communication primitives and device-side APIs that facilitate fine-grained, kernel-initiated data transfers across nodes and accelerators.
• Analyze and optimize GPU kernels comprehensively, pinpointing bottlenecks at the convergence of computation, memory, and networking, and driving focused enhancements.
• Work in tandem with network software, hardware, and AI framework teams to collaboratively design communication strategies that are compatible with GPU execution patterns and new model architectures.
• Develop proofs-of-concept, carry out experiments, and execute quantitative modeling to assess and verify new communication strategies prior to their production deployment.
• Aid in the advancement of programming models that reveal GPU-aware networking features to application developers.
• Over 5 years of practical experience in CUDA programming, including the writing and optimization of complex GPU kernels.
• A Master’s degree or equivalent experience in computer science, computer engineering, or a closely related discipline.
• A solid understanding of GPU architecture principles: warp scheduling, shared memory, L2 cache, memory coalescing, occupancy tuning, and asynchronous execution.
• Proficiency in systems-level C/C++ development within performance-sensitive environments.
• Knowledge of GPU data transfer mechanisms, such as GPUDirect RDMA and GPU-initiated communication.
• Capability to interpret and analyze GPU performance profiles (e.g., Nsight Compute, Nsight Systems) and convert insights into practical optimizations.
• Excellent collaboration abilities in a multinational, interdisciplinary setting.
• Health insurance
• 401(k) matching
• Flexible work hours
• Paid time off
• Remote work options
3Pillar Global
Stefanini Brasil
evoila
Get handpicked remote jobs straight to your inbox weekly.