Software Engineering Manager – NCCL

About the role

We are seeking a dynamic technical leader to head NVIDIA’s China NCCL team, responsible for delivering high‑performance communication libraries for deep‑learning and HPC workloads across thousands of GPUs.

Key responsibilities

• Lead, mentor and grow the China engineering team.
• Own end‑to‑end execution including planning, prioritization, quality control and performance.
• Engage with customers and researchers to capture use‑cases and requirements.
• Collaborate with engineering, program and product management to define the product roadmap.
• Contribute to feature design and implementation.
• Continuously improve processes, infrastructure and practices for efficiency and transparency.

Required profile

• 10+ years of software industry experience with at least 4 years in management.
• Bachelor’s, Master’s or Ph.D. in Computer Science, Computer/Electrical Engineering or related field.
• Specialization in systems software, communication runtimes or high‑performance networking.
• Proven success delivering complex initiatives through the full product lifecycle.
• Strong understanding of computer architecture, networking technologies and operating‑system fundamentals.
• Hands‑on C/C++ programming and Linux debugging expertise.
• Ability to balance multiple projects with competing priorities across time zones.

Required skills

• C/C++ development on Linux.
• RDMA, RoCE, Ethernet, InfiniBand, EFA networking.
• CUDA programming.
• MPI, SHMEM, OpenMP, OpenACC, pthreads.
• NCCL, NVSHMEM, UCX, UCC communication runtimes.
• Performance analysis and optimization of HW‑SW interactions.
• HPC and ML/DL fundamentals.
• Deep‑learning frameworks such as PyTorch, TensorFlow, vLLM, SGLang, TRT‑LLM.