Description
We are looking for a motivated Deep Learning engineer to bring advanced communication technologies into AI stacks, including PyTorch, TRT-LLM, vLLM, SGLang, JAX, etc. You will be working with the team that created communication libraries like NCCL, NVSHMEM & technology like GPUDirect -- for scaling Deep Learning and HPC applications. Your customers will have diverse multi-GPU demands, ranging from training on scales up to 100K GPUs to inference down at microsecond latency. Communication performance between the GPUs has a direct impact on AI applications. Your work in AI toolkits will make all of those easier for the community.
Responsibilities:
- Integrate new communication libraries features in AI frameworks: from PoC to performance analysis to production
- Perform deep analysis of AI workloads and frameworks to identify multi-GPU communication requirements and opportunities. Collaborate hands-on with teams working on the latest AI models.
- Improve AI compilers to hide communications or perform automatic fusion.
- Conduct in-depth AI workload performance characterization on multi-GPU clusters.
- Design fault-tolerant and elastic solutions for large-scale or dynamic AI workloads.
- Author custom communication or fused compute-communication kernels to showcase ultimate performance on NV platforms.
- Influence the roadmap of communication libraries - NCCL & NVSHMEM.
- Collaborate with a very dynamic team across multiple time zones.
Requirements:
- B.S, M.S. or PHD in Computer Science, or related field (or equivalent experience) with 5+ software engineering and HPC/AI experience
- Development or integration experience with Deep Learning Frameworks such PyTorch, JAX, and Inference Engines such as TRT-LLM, vLLM, SGLang
- Rapid prototyping and development with Python, C++, CUDA or related DSLs (Triton, cuTe)
- Solid grasp of AI models, parallelisms, and/or compiler technologies (e.g. torch.compile)
- Experience conducting performance benchmarking on AI clusters. Familiarity with at least one performance profiler toolchain (PyTorch profiler, NVIDIA Nsight Systems)
- Understanding of HPC/AI communication concepts (1-sided v 2-sided communication, elasticity, resiliency, topology discovery, etc)
- Adaptability and passion to learn new areas and tools
- Flexibility to work and communicate effectively across different teams and timezones
Nice to Have:
- Experience with parallel programming on at least one communication runtime (NCCL, NVSHMEM, MPI). Good understanding of computer system architecture, HW-SW interactions and operating systems principles (aka systems software fundamentals)
- Expertise in one or more of these areas: Training, Distributed inference, MoE, Reinforcement Learning, kernel authoring (on CUDA, Triton, cuTe, etc). Experience with programming for compute & communication overlap in distributed runtimes
- Experience with AI compiler pattern matching and lowering. Solid understanding of memory hierarchy, consistency model, and tensor layout
For more fresh new jobs, check out yubhub.co.