06 Aug
|
Technology next
|
India
06 Aug
Technology next
India
Senior GPU Systems Software Engineer
Job Type: Contract
Location: Remote (India)
Experience: 8+ Years
Job Overview
We are seeking a Senior GPU Systems Software Engineer with expertise in CUDA, NVIDIA GPU Architecture, Modern C++, and High-Performance Computing (HPC). The ideal candidate will design and optimize GPU-intensive applications for AI, Large Language Model (LLM) training & inference, scientific computing, and high-performance workloads.
Key Responsibilities
- Design, develop, and optimize high-performance CUDA kernels.
- Develop scalable GPU software for NVIDIA Ampere, Ada, Hopper, and Blackwell architectures.
- Optimize GPU performance through memory hierarchy, occupancy, warp scheduling, and instruction-level parallelism.
- Build multi-GPU applications using CUDA Streams, CUDA Graphs, and NCCL.
- Analyze and optimize PTX/SASS for maximum performance.
- Profile and debug GPU applications using Nsight Compute, Nsight Systems, CUPTI, and Compute Sanitizer.
- Collaborate with AI researchers to accelerate LLM training and inference.
Required Skills
- 8+ years of hands-on GPU programming experience.
- Strong expertise in CUDA and Modern C++ (C++17/20).
- Deep understanding of NVIDIA GPU Architecture, including SMs, Tensor Cores, Warp Scheduling, Memory Hierarchy, Shared Memory, and Occupancy Optimization.
- Experience with CUDA Runtime API, Driver API, Streams, Graphs, Cooperative Groups, Unified Memory, and Agile Parallelism.
- Strong knowledge of parallel algorithms, matrix operations, sparse computing, and tensor operations.
- Experience with NVIDIA Nsight Compute, Nsight Systems, cuda-gdb, CUPTI, and Compute Sanitizer.
- Experience optimizing multi-GPU workloads using NCCL.
Preferred Skills
- CUTLASS
- cuBLAS
- cuDNN
- TensorRT
- Triton
- Thrust
- OpenMP / OpenACC
- OpenCL
- Vulkan Compute
Preferred Domain Experience
- AI Infrastructure
- Large Language Models (LLMs)
- High-Performance Computing (HPC)
- Scientific Computing
- Deep Learning
- Machine Learning Acceleration
- Computer Graphics
- Physics Simulation
Pay: ₹90,000.00 - ₹150,000.00 per month
Experience:
- GPU: 8 years (Required)
Work Location: Remote
📌 Senior GPU Software Engineer (India)
🏢 Technology next
📍 India