NVIDIA CUDA Engineer

The NVIDIA CUDA Engineer is responsible for designing and optimizing GPU-accelerated software for high-performance, real-time environments. This role focuses on developing system-level software and parallel computation strategies across domains such as Embedded Systems, Artificial Intelligence/Machine Learning (AI/ML), and Quantum-Classical Hybrid Computing. Key Responsibilities: • GPU Programming: Develop and maintain CUDA-based algorithms; tune GPU kernels for maximum throughput and low-latency. • Systems Development: Build CUDA runtime libraries, drivers, and toolchain components; manage memory and performance profiling. • Heterogeneous Strategy: Implement parallel computation strategies across CPU/GPU/FPGA systems. • Collaboration: Partner with hardware architects to design holistic GPU solutions and accelerate AI workloads. • Quality Assurance: Enhance Continuous Integration/Continuous Deployment (CI/CD) pipelines and benchmark software performance. Essential Skills: • Expert-level C/C++ programming. • Deep expertise in NVIDIA CUDA and GPU Architecture. • Proficiency in Performance Tuning and profiling tools (e.g., Nsight, nvprof). • Strong grasp of Parallel Programming Paradigms (Multi-threading, Vectorization). • Experience in System-level Software and Memory Management. Desirable Skills: • Experience with Compiler Design (LLVM, MLIR). • Knowledge of Quantum-Classical Systems (CUDA-Q). • Background in Real-time Algorithms and AI Model Acceleration. • Master’s degree in Computer Science or Electrical/Electronics Engineering.